Rail-rail opamp, low low noise, 25v or so?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Svart

Well-known member
Joined
Jun 4, 2004
Messages
5,134
Location
Atlanta GA USA
Anyone have a favorite opamp that can do rail to rail around 25v? I'm looking for something with as low noise(V and I) as possible..

I'm playing around with wideband VCOs and I need as many different opamps to try as level shifters. The PLL I am using only goes from 0-5vdc but the only options I have for VCOs seem to be 0-20v(with the bandwidth that I want to use).

I am using an LT1038 right now but it only gets within 2v of the rails(ground and VCC) and I don't want to end up biasing it with a negative rail if I don't have to.

Another idea is to use some kind of discrete transistor but I really don't want to spend an excessive amount of time getting that right... Any opinions?
 
A FAE from Microchip was trying to push some of their rail to rail opamps on me the other day.. I can look up the part numbers if ur interested. I think they were CMOS so probably not very low noise, but as I recall R to R.

He said they were low noise, but that's his job.

JR
 
yes please forward me the models. I'll take a look at their website to see if there is any info there.

there are tons of rail-rail opamps, but it seems that once you get above about 5-8v rated parts the fewer there are.
 
I guess you mean rail-to-rail output, only, right?

I think you are right though that the r-to-r parts are virtually all low voltage. To do what you want I would probably roll my own, a few discretes along with a decent IC opamp driving. If the load Z is not very high you might manage to buffer the output with a 'high-voltage" CMOS gate and figure out the compensation. Wouldn't be production-worthy since the d.c. parameters of those gates are not guaranteed. And this would only get you to a max swing of 18V (and that is pushing it). Probably a bad idea.

A grounded emitter with low-or-no emitter R and a diode-connected Q as a simple current mirror on one rail, and a comparable stage on the other rail, or just a current source there, might not be too much of a bear to compensate. How fast does the performance in-circuit have to be?
 
MCP6021 (or 2 or 4) which is 10MHz and 9nV/rtHz
--------
My bad... it looks like a 5V process...CMOS is puny.

It shouldn't be that hard to roll your own with a transistor or two.

Or flip the problem and just run the opamp off a wider supply rail.


JR
 
The goal is to get 0-20v range for the VCO from the 0-5v PLL loop out. I'm trying to figure out a reliable way to get that from an opamp that is biased from ground to VCC with no negative rail. It might not be possible but I have to try.. :wink:

So that would make it rail-rail output

As for speed.. the who thing has to be able to jump to any frequency within the VCO bandwidth and lock/settle within a 1.5ms window. I'd prefer the whole thing to be within 1ms. To compound things I have to make a SLM reading during that period too! So I can't deal with a lot of phase noise, V/I noise or simple jitter since I don't have a lot of time for things to settle down.

To give you an idea of what I am doing, this device will be running a test routine that tunes the tuner to a specific frequency every 2ms. Once it acquires a dummy signal at that frequency, both the source and the receiver(the part I'm working on) hop to the next frequency and so on.

If you haven't guessed, it's pretty much a sampled IF SDR receiver that we are trying to simplify enough to be cost effective in a consumer commercial/test and measurement market.
 
VCO input drive is not very well specified in the datasheet:

http://www.minicircuits.com/pdfs/ROS-2250W-119+.pdf

However we use a number of their parts and we should easily be able to squeeze the info out of them.. :green:
 
http://www.analog.com/UploadedFiles/Data_Sheets/OP184_284_484.pdf

gets within 125mv of the negative rail.. Not sure how it would work to ground but I'll assume that it is similar.

The VCO's minimum frequency is stated at 1220mhz but that is roughly at 800mv..

I think that could work but the noise is fairly high compared to what the RF guys are suggesting. I suppose a couple of samples can't hurt.
 
Interesting part. At the speed you have described you could probably lowpass filter the output a little and still be able to move quickly enough. Already 3.9nV/sq rt Hz is not bad at all. Any feedback components could easily add more noise than that unless they are pretty low Z.
 
[quote author="mikep"]the one that comes immediately to my mind is OP777. R-R output and IIRC 30V max.

mike p[/quote]

That also looks interesting, although a lot slower than the 184 series and a good deal higher voltage noise (15nV/sq rt Hz).

http://www.analog.com/UploadedFiles/Data_Sheets/OP777_727_747.pdf
 
[quote author="Svart"]VCO input drive is not very well specified in the datasheet:

http://www.minicircuits.com/pdfs/ROS-2250W-119+.pdf

However we use a number of their parts and we should easily be able to squeeze the info out of them.. :green:[/quote]

Looking at that datasheet, the fact that the abs max tuning voltage is so much more than the abs max supply voltage sort of suggests that the current drive requirements are likely low---as if you were hooking up pretty much directly to a varactor diode or pair of same, which are going to generally be reverse-biased. Now, they are likely going to put some resistor there to reference the diode to something when the control voltage terminal is open-circuited. If you made enough of these you could probably persuade them to leave that out, or at least make it a lot bigger, if its loading is significant.

Another question: how accurate and stable does this voltage translation operation need to be? Since noise has been mentioned as a consideration maybe it is a pretty stringent requirement.
 
A few brief ramblings:

- You only need to be able to pull down to (close to) the negative rail; positive compliance can be produces by upping the positive rail.

- this appnote may include relevant information.

- as you're effectively building an SDR spectrum analyzer, why not have a (controlled) ramp at the VCO input and correct for it either through a DAC or a counter? Those Mini-Circuits VCOs are pretty predictable/repeatable, and a simple linear/quadratic interpolation will likely end up being close enough.

JDB.
 
Another question: how accurate and stable does this voltage translation operation need to be? Since noise has been mentioned as a consideration maybe it is a pretty stringent requirement.

Some of those specs haven't been totally hashed out yet so I'm proceeding under the premise that "the best I can get" is my best route until those above me decide to get their act together.

jdbakker, That appnote is great. I found that the LT1006 mentioned in the note is a very good fit as well as showing me that the BJT level shifter idea that I had is also a viable alternative.

Thanks all, I'll keep you posted.
 
Also, to those interested:

It works.. Like a champ. I just have to clean up some phase noise and recalculate the loop BW filter for the new opamps and I'm on to the next project.

:thumb:
 
That 8675 looks like a honey in many ways. A little pricey for some applications though.

EDIT: Looking now at the datasheet in detail, slew rate is a little low and current noise a bit high---the latter because it is probably a bipolar input with bias current compensation. Still in all, a good amp for d.c.-to-moderate frequencies and low-medium source impedances.
 
AD820 is excellent for onboard instrument preamps.
What about a LT1007?
Then there's OP275 ...
What about TLE2071 they're cheap?
 

Latest posts

Back
Top