Implementing a “Y Out” on an Oscilloscope

[wintermute]

Half your luck George, I think a lot of diy'ers would love to have your form of "rent" for use of equipment when doing their diying!

[gpapag]

Originally Posted by wintermute

Half your luck George, I think a lot of diy'ers would love to have your form of "rent" for use of equipment when doing their diying!

I don't know.
But if they'd love to, is it so difficult for them to p*** their wifes off with their "men childish activities"?
(That's what I do. Not on purpose of course)

Edit:"men childish and useless(i.e. diyaudio) activities" buhaha

[gpapag]

To continue with the manual bandwidth measurements of the 1:1 bifilar x-former.
Set-up as per Measuring the input impedance, bandwidth and efficiency of transformers

With 1Vrms at the primary the –6dB LF was ~30Hz.
Oops that’s too high. Will I have to bake another one?

I repeated the manual measurement, this time varying the amplitude of the input signal. (att.1)
There is a faint pattern coming out. The lower the signal, the higher the LF cut-off point. This is somehow the opposite from what I was expecting to happen.

I set-up an automatic measurement (att. 2) to have a more consistent picture of this pattern.

Results in att.3. (recording/analysis SW: REW V5.00)

Shown is Frequency response of the 1:1 bifilar with varying the input signal.
This is achieved by adjusting the input sensitivity of the oscilloscope and then by adjusting the output of the soundcard to get to the lowest signals.

LF roll-off is confirmed.

Numerical details are patched on, which leads to att.4

Regards
George

[wintermute]

LOL so when are you starting a transformer winding business George

I might have to look at this again tomorrow, I'm a bit tired (and have had a couple of glasses of wine) so it isn't sinking in... I note that you also seem to have a problem with 50Hz interference like I do though... I got some clean measurements early on with none, but can't seem to get them any more, I don't know what changed

Tony.

[gpapag]

LOL so when are you starting a transformer winding business George

When???
I am in full production if you did't notice!

Tony, I suggest you would better go to bed. The boss is waiting.
And mind your activities there, you had a couple of reds already.

There are a few things that worth discussing on the graph. But I am a bit tired too. I'll try in a few hours.

Regards
George

[gpapag]

So what is this diagram with the so many curves on it?

This is basically a Signal Amplitude versus Frequency screen, upon which many frequency response curves have been dumped.
Horizontal axis is frequency (Hz)
Vertical axis is in dB. As shown in the attachment of the previous post, the range was arbitrary set by the software.
Now I edited the numbers to reflect the reality of the measuring set-up.

Therefore, on Vertical axis, 0 db is on the top and progresses downward as –db.
0dB is the highest signal that can be inputted in the sound card.

It is important to specify this 0dB signal.

How I did it? By applying continuously a sinusoidal signal around 1kHz at the card’s input. While watching the VU or Peak meters of the recording software, I adjusted the signal generator’s output, till the Vu meters showed exactly 0dB. Using an AC rms voltmeter I measured the signal entering the Sound Card.

0dB with my card equals 1.286Vrms.

After establishing the true amplitude of 0dB it is feasible to translate in Vrms any sinusoidal signal that produces a certain VU meter indication.

This way, at the right side end of each curve, it is written the signal amplitude in dBs relative to 0dB of my card, as well as the corresponding real value in mVrms.

And this is important for me because I easily forget how weak can a signal be e.g a humble -48db equals 4mV.

At the left side in the area of 20Hz, I have marked on each curve how many dbs rel to 3kHz is this curve down at 20Hz.

As you can see, with this x-former, the smaller the signal, the heavier the drop in LF, (confirmation of the manual bandwidth measurement results).

Can anyone provide a possible explanation for this?

This diagram shows me also how vulnerable can small amplitude signals be to extraneous noise and how important is to properly plan, arrange for and implement a measurement. Here, I have to quote SY

Originally Posted by SY

The sad reality of high sensitivity measurement is that it takes almost no time to do the measurement, but a LONG time to set things up properly!

Regards
George

[wintermute]

Hi george, have you done a loopback measurement on the sound card with the varying drive levels, to check whether or not the difference is the card's frequency response, as opposed to the transformers?

Thanks for the explanation that helps a lot

I had a thought this morning, with respect to the 50Hz problem. I've had similar problems in the past (worse than my current ones) due to earth loops with the PC, and I suddenly thought: "If I had a high quality usb soundcard, I could do all of the measurements on my laptop running off battery, That should completely eliminate any 50 Hz that is coming from the PC, and also will remove the possibility of an earth loop. I might need to revise my Birthday present from a capacitor tester (in circuit ESR meter) to a good USB sound card

Tony.

[gpapag]

Oops
I spoted a small mistake on the diagram.
I had marked the odd harmonics of main's 50Hz, but missed the 7th.
Now I corrected it. Sorry

Hi george, have you done a loopback measurement on the sound card with the varying drive levels, to check whether or not the difference is the card's frequency response, as opposed to the transformers?

Tony I will do it and report back. Thank you.

May we all have a good and fruitfull new month (June)

Regards
George

[gpapag]

Originally Posted by wintermute

Hi george, have you done a loopback measurement on the sound card with the varying drive levels, to check whether or not the difference is the card's frequency response, as opposed to the transformers?

At your service sir

Tony, as you see, the card’s response is ruler flat.
Signal ranges from +3dB (yes, I overloaded the input a bit to see how it behaved. In this card, the inputs are overload protected by clamp diodes) to –78dB

Note: This is the response of the measuring RH Input channel, which was connected directly to the LH Out .

The LH Input was left open, so there is no compensation or correction of any kind.
This is why the noise from the shorted RH Input is around –90 at 20kHz.

If the LH Input too was connected to LH Out, as is usually done for to make the LH Input channel the “reference”, then the shorted RH Input would be another 30 to 40 dBs lower (around –130dB) at 20kHz.

Regards
George

[wintermute]

Thanks George, Was worth checking! I spent a heap of time trying to work out why the freq response of my bass reflex enclosure was no where near the sim, it turned out it was because my microphone was damaged (after dropping it). Tried a new mic and the nearfield matched the sim almost perfectly!

The effect is quite subtle, I have no ideas as to what might be causing it.

Tony.

Tony.