Op amps and component tolerance

Thread Starter

geoffers

Joined Oct 25, 2010
320
Hi all,

If built a few devices based around the 16f1779, using the onboard op amps. I realise its not the best spec op amp and while they all work, I've noticed a variation in performance.

Its a RFID reading device and I'm happy with how well it works but might make more that need better range.

I've used the cheapest surface mount resistors and capacitors, would folks think spending a bit more on discrete components with better tolerance is the way forward or should I be looking at better spec op amps?

Cheers Geoff
 

Delta prime

Joined Nov 15, 2019
337
would folks think spending a bit more on discrete components with better tolerance is the way forward or should I be looking at better spec op amps?
Their is a very popular general purpose op amp .The 741 operational amplifier.(Audio guru again's favorite).He is a very distinguished Member of AAC .:p
 

Papabravo

Joined Feb 24, 2006
13,759
Hi all,

If built a few devices based around the 16f1779, using the onboard op amps. I realise its not the best spec op amp and while they all work, I've noticed a variation in performance.

Its a RFID reading device and I'm happy with how well it works but might make more that need better range.

I've used the cheapest surface mount resistors and capacitors, would folks think spending a bit more on discrete components with better tolerance is the way forward or should I be looking at better spec op amps?

Cheers Geoff
This is a tough one. The parameters of a semiconductor opamp are difficult to control with precision and repeatability. It was recognized back in the 1930's that the ability to design an amplifier whose properties depended on the passive components was a great advantage. The magic bullet was FEEDBACK. The properties of surface mount resistors can be controlled with extraordinary precision; capacitors not quite so much but better now than they used to be. The caveat in all of this is that it really depends on what you are doing. Tell us more about what the analog configurations you are using are doing and maybe we can be more helpful.
 

Thread Starter

geoffers

Joined Oct 25, 2010
320
Thanks for the replies,

The carrier is 137 kHz and using fdxb and hdx protocol.

I have some PCBs that I put a lot of time to designing and I'm happy with how well they work, but between boards I see a difference, some read the hdx better some the fdx, I guess this is coming from the discrete components tolerance?

What I'm really asking where should the money go first if I want to improve things? Discrete components makes sense as I have the design already. But if the op amp is the restriction a redesign may be needed.

My first design used a tl082's and once I got rough values I worked back to use the onboard pic op amp, longer range readers are pretty expensive so I assume more expensive op amps are used?

Cheers Geoff
 

Papabravo

Joined Feb 24, 2006
13,759
Thanks for the replies,

The carrier is 137 kHz and using fdxb and hdx protocol.

I have some PCBs that I put a lot of time to designing and I'm happy with how well they work, but between boards I see a difference, some read the hdx better some the fdx, I guess this is coming from the discrete components tolerance?

What I'm really asking where should the money go first if I want to improve things? Discrete components makes sense as I have the design already. But if the op amp is the restriction a redesign may be needed.

My first design used a tl082's and once I got rough values I worked back to use the onboard pic op amp, longer range readers are pretty expensive so I assume more expensive op amps are used?

Cheers Geoff
Anecdotal evidence is useful as a guide to further research, but there is no substitute for analysis and experiment. If I were you I'd dig out one of the prototype boards with the external opamp and do some side by side comparisons until you can identify the factors which affect the size of the reliable capture range. It would be handy if you had multiple units to do this with. Would you have to see NDA's for you to share a schematic of just the analog part with us. I'm curious about things like amplifier gain, frequency response, phase response. If you are just using the opamp as a buffer, then maybe there is something else.

Off the top of my head in favor of the TL082 I see:
  1. Wide Vcc range ±15 VDC with some headroom to absolute Max
  2. Unity gain bandwidth is 3 MHz.
  3. Slew rate is 13 V/μsec.
I'm guessing that the onboard device is operating from a single supply. This may or may not be a limiting factor.
 

kubeek

Joined Sep 20, 2005
5,724
What I'm really asking where should the money go first if I want to improve things? Discrete components makes sense as I have the design already. But if the op amp is the restriction a redesign may be needed.
My first design used a tl082's and once I got rough values I worked back to use the onboard pic op amp, longer range readers are pretty expensive so I assume more expensive op amps are used?
As I said, without at least a partial schematic there is little to talk about. TL082 and similar have quite a small GBW of about 3Mhz, so if you have receive gain over about 20 I might be inclined to say that variations in the opamp and if it exceeds that spec or not could be the culprit of your differences. To appease the spirit of Audioguru, I tend to put TL06x,07x,08x parts in the same box as LM741 - old, slow, and not really good for anything other than showing students all the possible shortcomings an opamp could have.
So please show a diagram so we have something concrete to talk about.
 

Thread Starter

geoffers

Joined Oct 25, 2010
320
Thanks, I'll try and pull the analog bit off the schematic in the next couple days, just looked at it and its pretty busy, the op amp bit though is fairly simple with just a few discrete components.

It is single supply which I'd like to keep for this design but would go dual if I wanted to make a improved reader.

Cheers Geoff
 

MisterBill2

Joined Jan 23, 2018
6,107
Hi all,

If built a few devices based around the 16f1779, using the onboard op amps. I realise its not the best spec op amp and while they all work, I've noticed a variation in performance.

Its a RFID reading device and I'm happy with how well it works but might make more that need better range.

I've used the cheapest surface mount resistors and capacitors, would folks think spending a bit more on discrete components with better tolerance is the way forward or should I be looking at better spec op amps?

Cheers Geoff
There is generally a published tolerance specification for resistors and capacitors and if the parts that you get are within those specifications then that should not be the issue. Adding an external amplifier IC is certainly a way to have more gain, and the frequency response of those op-amps will certainly matter a great deal. Thus a study of the data sheets and the manufacturers recommendations is a requirement. You will not find any of the needed data at the less reputable sellers websites, though.
 
Top