OK, some of you may have read the thread on Reflections and Attenuators, and will recall I used a DAC, that had a
horrible input return loss. This was caused by the 2 huge ferrite beads, in series with the transformer primary.
Let's not get into whether ferrites are evil, or not. A lot of companies stick them in the SPDIF output, as it will limit the amount of HF energy, that it spews out. Sure, horrible for sound. But, those companies are a lot more interested in meeting regulatory standards, than placating audiophiles.
Having said that...........
I can not see much of a reason for using them on the SPDIF input. Well, yeah, the same reason, so let's see how much crud is really there. And whether ripping them out helps, or not.
So, assuming you have a mental image of what the stock version was, here is what it looks like, with the ferrites replaced with wire:
I need to point out the vertical scale is 1/2 half of the previous pictures. Since we are going to do a lot of tweaking, the sensitivity needs to be more.
OK, three things should stand out:
1.) The huge spike is gone.
2.) The impedance of the input stage is lower than 75R. (More about this, in a minute.)
3.) There is some crud, about 1 div, to the right of the end of the decay.
That is only there, when the unit is on, and the SPDIF input is selected. With one of the other inputs selected, the crud goes away, and the impedance is 75R.
OK, that #2............
Yes, the circuitry
does load the impedance. Sure, not much, but it does. So, to the experts who are 100% certain that all you need is a 75R termination.............yeah, wrong. Again.
So, how much is it off?
I removed the 75R resistor, and stuck a trimpot, in its place. Adjusted it, and got this:
The trimpot turns out to be
80 ohms. So, there!
OK, at this point someone may notice the reflection hump has moved.
Yes, it has! I changed cables, to one that I am certain is 75R. Just to make sure I am not trying to match to a cable that is really 78R. And it is longer. So, 2 cables, both show the DAC is lower than 75R. I'm sure the peanut gallery will still be claiming I am full of it, and the cables are wrong. Whatever..........suit yourself. Just don't interrupt the thread. Please. OK?
Ok, this actually looks pretty good. Even with a yucky RCA. (And we all know how much I hate RCAs.) But, can we do better?
Sure!
"How?"Well, at this point, I am going to be evasive, and claim "trade secret". But, I will give you enough, to come close. (I have to point out 98% of my income is derived from the consulting side, of the business. The manufacturing side [ar-t] isn't where I earn a living. So, I can't give away the farm. I charge $$ for coming up with the right way to do this. So, the equation is not something I am going to share. Sorry.)
OK, what am I getting at.....................termination.....
..........
Everyone thinks all you needs is 75R. Well, already showed that is wrong. Assuming one figures out that resistor may need to be a bit higher, you are closer, but still wrong.
Just because the data sheet says it will work with 75R, does not mean it works best, with 75R.So, what needs to be done.............is............is......
.....ah..............ya gotta fudge things, quite a bit.
What kind of fudging, and how much? Ah, that is the secret. But, the crux is every transformer works best with a certain load impedance. (How that is calculated...........TDR and years of experience.) But, I'll try to 'splain it, in a manner that you guys can have fun tinkering with.
Let's say that transformer works best with 300R load. So, you would put 300R, on the secondary, and stick 100R, on the input, to get 75R. Or, 100R on the secondary, and 300R, on the primary.
Or 150R, on both.
Get the gist?
Ok, so having come up with a guess, as to what they should be, here it is:
Ya know.................RCA, adaptor............all that, and the leakage inductance, of the transformer.............it looks pretty good. (Yes, if I was at the lab, and not my garage, had some BNCs, and the parts to make a zobel, it would look better. But, 99% of you guys don't have close to that much test equipment, so here is something you can do.)
Ok, you probably want to know how it looks, with the original vertical sensitivity.
I would say that looks good, all things considered. Signal is around 5 div.............reflection is a bit less than 1/2 div, so we will say the rho is around 0.1. Or, -20 dB. With an RCA, and very little tweaking, not bad.
Compare that to what we started at..................
OK, all for now. Asbestos suit on.
Pat
