I understand what you are doing but I think you are not using the term reciprocity correctly, just a detail not a criticism. The temperature to resistance relationship can vary between like devices, it is not fundamental.
@DNi - Post 5428 clarified it...and your follow up. Thanks.
Yes PM sent.
When you explained the importance at low frequencies, it made sense...
because if the fundamental shifts..slightly out of bin, then it is a series of
measurements moving about the fundamental frequency's axis...(for lack of a better term).
In short, it gets tricky doesn't it?
And at extremely low levels approaching SOTA...
...tricky doesn't even begin to describe the challenge.
I don't feel so bad now....that with the 5 year old thread
there is still a lively technical discussion with twits and
turns about and lots of investigating and testing, etc.
This is all very good stuff.
Funny though Davada ended up with a lot more opamps
that I did...how did that happen? And that fancy resistor
to boot!
Life goes on....
Cheers,
Sync
Yes PM sent.
When you explained the importance at low frequencies, it made sense...
because if the fundamental shifts..slightly out of bin, then it is a series of
measurements moving about the fundamental frequency's axis...(for lack of a better term).
In short, it gets tricky doesn't it?
And at extremely low levels approaching SOTA...
...tricky doesn't even begin to describe the challenge.
I don't feel so bad now....that with the 5 year old thread
there is still a lively technical discussion with twits and
turns about and lots of investigating and testing, etc.
This is all very good stuff.
Funny though Davada ended up with a lot more opamps
that I did...how did that happen? And that fancy resistor
to boot!
Life goes on....
Cheers,
Sync
Last edited:
Thank you Scott for clarifying this. Now I understand a little better why
we are using 997 Hz, that Demian used and mentioned that we should also
use it, but never explained why.
I was curious and wondered why we wouldn't use the more standard
1004 Hz that Ma Bell (referenced to the original Bell Labs, Western Electric,
and ATT [the old one]) used to use and all the smaller systems.
With the advent of digital I guess the **** from 1004Hz to 997 makes sense
in the digital world. Just as long as we don't get into those really long prime
numbers multiplied to each other....We'd start getting "visits".
Cheers,
Sync
As usual... We used to not agree on the meaning of nonohmic.
Yes there are resistors that have multiple significant contributors to resistance variations such as wire wound types. But as we get into the higher quality types they get much more uniform.
The 4/1 1/4 bridge technique has also allowed me to look at capacitors. But unlike resistors using a bridge to match devices first allows a decent test.
But I wouldn't try it on inductors. Too many other variables.
Currently working on cable. Seems to be interesting characteristics that should be obvious but for whatever reason ignored.
Last edited:
Thanks, neither Mouser nor Digikey know of this part number yet.
"Turn left where they plan to build a new bank"
But the low supplies provide an excuse to ask about rail-to-rail outputs .
Anyone here know of any references on how to optimize these?
Op-amp factory people must have looked at this but I can't find much.
Maybe someone on this thread
Best wishes
David
Thanks, neither Mouser nor Digikey know of this part number yet.
"Turn left where they plan to build a new bank"
But the low supplies provide an excuse to ask about rail-to-rail outputs .
Anyone here know of any references on how to optimize these?
Op-amp factory people must have looked at this but I can't find much.
Maybe someone on this thread
Best wishes
David
Linear - Amplifiers - Video Amps and Modules | Integrated Circuits (ICs) | DigiKey
Back in the old days...er the Wiki works days, we could try and transfer the knowledge
to the wiki. With links back to the source. But it got broke and now no one can find
anything on/in the wiki. Over managing, micro-managing, trying to force free thought down one narrow path isn't great.
Trying to let people think freely and express freely while challenging isn't that bad.
Just remember we can always put a TOC on the first page of a thread as it grows
with important markers/links for those who come along later.
Things to think about,
Cheers,
Sync
to the wiki. With links back to the source. But it got broke and now no one can find
anything on/in the wiki. Over managing, micro-managing, trying to force free thought down one narrow path isn't great.
Trying to let people think freely and express freely while challenging isn't that bad.
Just remember we can always put a TOC on the first page of a thread as it grows
with important markers/links for those who come along later.
Things to think about,
Cheers,
Sync
Speaking of thread moderation:
I found this amusing. What I believe happened was an attempt at typing a five-letter word with an f in it, but the f got dropped, and the bbs software recognized it as a word to be bleeped out. I always thought such auto-bleeping software was shifty.
I like the TOC idea, though someone, either the original poster or a moderator, has to edit that first post to add that info (and if this is done, please use post number instead of page number - I read at 50 posts per page, and some use the default of 20 posts per page, so page numbers can be confusing).
But this seems to be the same problem as the wiki, it requires organization, and someone willing to take the time to put into doing it.
What are the hidden issues? I always thought the sin^2 + cos^2 detector to be elegant, as it gives a continuous and instantaneous signal level for an oscillator that has quadrature output. I recall posting early in this thread about using a microcontroller to do this, though two analog multipliers would work as well. As I recall, the main concern was reducing the long settling times at low oscillator frequencies when sampling or a diode peak level detector is used.
Hi Ben,
The software does delete some words completely automatically. Keep in mind that some words can be offensive in different societies,and we try to keep the worst of it out. Don't even try to guess at the word as I have no idea what it really was either.
I don't see what you think is "shifty" about it. It is a gentle reminder to try and keep your posting at a professional level. To each his own ...
-Chris
Edit:
The software does delete some words completely automatically. Keep in mind that some words can be offensive in different societies,and we try to keep the worst of it out. Don't even try to guess at the word as I have no idea what it really was either.
I don't see what you think is "shifty" about it. It is a gentle reminder to try and keep your posting at a professional level. To each his own ...
-Chris
Edit:
... and who would maintain these TOCs?
Available analog multipliers do not offer enough performance. If you operate them at low level, the control voltage will be noisy which shows up as oscillator sidebands. If you operate them at high level, large-signal defects (nonlinearity) result in inacceptable ripple components, which as we know cause distortion.
Also the approach depends on the level of the quadrature outputs to be identical. For this the time constants of the two integrators need to match, which is highly inconvenient to trim.
You could use ADCs to digitize the signal and do the math in a DSP (perhaps also solving the issue with the matching time constants), but if this ends up cheaper than a S & H-based approach would need to be determined.
Samuel
Hi David,
If it isn't unaccounted for effects, then the math isn't right. I'd have to say that at the levels you are working at, the variables are the issue.
-Chris
Well, it really does. It's just that there are too many variables in effect. Sometimes the darnedest things can be important.
If it isn't unaccounted for effects, then the math isn't right.
I'd have to say that at the levels you are working at, the variables are the issue.-Chris
Hi David,
Well, it really does. It's just that there are too many variables in effect. Sometimes the darnedest things can be important.
If it isn't unaccounted for effects, then the math isn't right.
-Chris
It's not the math it's the models, you can tell your simulator that your FET is perfectly square law but you can't make one that is.