Author Topic: Diffrential amp design  (Read 5693 times)

0 Members and 1 Guest are viewing this topic.

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Diffrential amp design
« on: January 23, 2020, 07:30:41 am »
Hi,
I have an old design and it's working perfectly, it's used in an Electromagnetic Flow Meter, the input sensor differential potentials is in uV ranges.

Now I want to redesign it with newer and cheaper parts, because I want to increase the performance, and use lower priced parts, also It would be very beneficial if i could do it in a single supply config and also If I could do it fully differential and remove the GND from the input sensor, currently I'm using AD620 part, But for the new design I prefer using normal op-amps or use Chinese parts! any Ideas are highly appreciated.  If I could do it under 1$ for total op-amps it would be perfect.
ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Online moffy

  • Super Contributor
  • ***
  • Posts: 2217
  • Country: au
Re: Diffrential amp design
« Reply #1 on: January 23, 2020, 08:10:17 am »
Because of the micro volt signal level, I  would go with a true differential amp as your AD620, or risk getting swamped by low cmrr. But you don't need two, the second can be just a straight opamp.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20360
  • Country: gb
  • 0999
Re: Diffrential amp design
« Reply #2 on: January 23, 2020, 08:38:43 am »
Because of the micro volt signal level, I  would go with a true differential amp as your AD620, or risk getting swamped by low cmrr. But you don't need two, the second can be just a straight opamp.
Yes, it's not possible to do this with a cheap op-amp, not without spending more on precision resistors. Accurately measuring microvolt signals is incompatible with low cost.
 

Online magic

  • Super Contributor
  • ***
  • Posts: 7453
  • Country: pl
Re: Diffrential amp design
« Reply #3 on: January 23, 2020, 08:43:14 am »
You can eat your cake and have it too: buy AD620 from AliExpress 8)
 

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Re: Diffrential amp design
« Reply #4 on: January 23, 2020, 08:57:32 am »
 ::) :P :-[ :-X we have made a lot progress in the industry, why stick to some old and pricey tech!
Can a few these  zero-drift Precision OpAmps come in handy? Like GS8333
https://lcsc.com/product-detail/Precision-OpAmps_Gainsil-GS8333-TR_C431326.html
ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Online moffy

  • Super Contributor
  • ***
  • Posts: 2217
  • Country: au
Re: Diffrential amp design
« Reply #5 on: January 23, 2020, 09:37:30 am »
::) :P :-[ :-X we have made a lot progress in the industry, why stick to some old and pricey tech!
Can a few these  zero-drift Precision OpAmps come in handy? Like GS8333
https://lcsc.com/product-detail/Precision-OpAmps_Gainsil-GS8333-TR_C431326.html

The problem is not the opamp, but if you do it without a diff amp it's the precision and cost of the resistors over temperature to get your CMRR to good levels. They tend to be more expensive than the true diff amp.
 

Offline OwO

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: Diffrential amp design
« Reply #6 on: January 23, 2020, 09:53:42 am »
I looked for cheap instrumentation amplifiers a while ago and didn't find any. There are probably 10 different Chinese op amp manufacturers and you can find everything from 500MHz op-amps to low offset op-amps, but it seems no one has ever heard of instrumentation amplifiers. I think it's probably because trimming the resistors is expensive. Some of the datasheets even suggest you build an instrumentation amplifier using op-amps and discrete parts. It depends what kind of price break you are getting on the AD620. If it's close to digikey prices then it's still far cheaper to do it with discrete op-amps and 0.1% resistors. Do the most obvious thing and keep all of the gain in the first stage. This way your total CMRR is the second stage CMRR multiplied by the gain.
« Last Edit: January 23, 2020, 09:56:21 am by OwO »
Email: OwOwOwOwO123@outlook.com
 

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Re: Diffrential amp design
« Reply #7 on: January 23, 2020, 10:10:37 am »
Quote
The problem is not the opamp, but if you do it without a diff amp it's the precision and cost of the resistors over temperature to get your CMRR to good levels. They tend to be more expensive than the true diff amp.
0.1% resistors are super cheap right now

Some values for reference
https://lcsc.com/product-detail/Others_Viking-Tech-AR03BTCX1000_C319953.html
https://lcsc.com/product-detail/High-Precision-Low-TCR-SMD-Resistors_100KR-1003-0-1_C122538.html
ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Offline Warhawk

  • Frequent Contributor
  • **
  • Posts: 836
  • Country: 00
    • Personal resume
Re: Diffrential amp design
« Reply #8 on: January 23, 2020, 10:32:09 am »
You can eat your cake and have it too: buy AD620 from AliExpress 8)
Yeah, 100% genuine

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20360
  • Country: gb
  • 0999
Re: Diffrential amp design
« Reply #9 on: January 23, 2020, 12:14:44 pm »
Quote
The problem is not the opamp, but if you do it without a diff amp it's the precision and cost of the resistors over temperature to get your CMRR to good levels. They tend to be more expensive than the true diff amp.
0.1% resistors are super cheap right now

Some values for reference
https://lcsc.com/product-detail/Others_Viking-Tech-AR03BTCX1000_C319953.html
https://lcsc.com/product-detail/High-Precision-Low-TCR-SMD-Resistors_100KR-1003-0-1_C122538.html
No, that's not good enough. The AD620 has a CMMR of 130dB, which is much better than you'll get with 0.1% resistors.

The circuit you posted at the start of the thread is nearly twice as expensive as  necessary. The second AD620 is being used as a non-inverting amplifier and should be replaced with ordinary op-amp. Keep the AD629 for the first stage and increase the gain so the second stage can be unity gain. This will both improve the performance and reduce the cost.

To reduce cost further, you might be able to omit the second stage, by reducing the impedance of the first filter, although that would require larger capacitors, which cost more. There's also no point in using 1% E96 resistor values such as 169k and 30k1 for the filter, because the capacitors probably have a wider tolerance than 5%. Use the nearest E24 values: 160k and 30k.

EDIT: The total gain of your circuit is 71.3k so it will need more than one amplifier. Two amplifiers with a gain of roughly squareroot of the total required gain, will do.
« Last Edit: January 24, 2020, 09:24:05 am by Zero999 »
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Re: Diffrential amp design
« Reply #10 on: January 23, 2020, 02:08:16 pm »
You can eat your cake and have it too: buy AD620 from AliExpress 8)
Yeah, 100% genuine

 ;D
 

Offline technix

  • Super Contributor
  • ***
  • Posts: 3508
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
Re: Diffrential amp design
« Reply #11 on: January 23, 2020, 05:19:36 pm »
There might be some ADC with true differential inputs and has enough resolution natively. If that is the case, scratch the whole front end and have your sensor hooked directly to that ADC and call it a day. Maybe put a can over that, but anyway use as short an analog section as possible.

As long as the noise is under control, systematic biases and nonlinearity can be trimmed out afterwards in software. You don't even need to know exactly what kind of nonlinearity you may have, just assume the inverse transfer function is a polynomial (actually the first few terms of the function's Taylor series, but that is usually enough since you have a limited range anyway) and grab a few calibration points so you can work out the coefficients.
« Last Edit: January 23, 2020, 05:22:54 pm by technix »
 

Online Marco

  • Super Contributor
  • ***
  • Posts: 7044
  • Country: nl
Re: Diffrential amp design
« Reply #12 on: January 23, 2020, 07:01:37 pm »
I looked for cheap instrumentation amplifiers a while ago and didn't find any. There are probably 10 different Chinese op amp manufacturers and you can find everything from 500MHz op-amps to low offset op-amps, but it seems no one has ever heard of instrumentation amplifiers. I think it's probably because trimming the resistors is expensive.
That's no excuse, the transconductance based differential amplifiers like MAX4208 don't rely on resistor matching.
 

Offline MasterT

  • Frequent Contributor
  • **
  • !
  • Posts: 851
  • Country: ca
Re: Diffrential amp design
« Reply #13 on: January 23, 2020, 07:20:04 pm »
I looked for cheap instrumentation amplifiers a while ago and didn't find any. There are probably 10 different Chinese op amp manufacturers and you can find everything from 500MHz op-amps to low offset op-amps, but it seems no one has ever heard of instrumentation amplifiers. I think it's probably because trimming the resistors is expensive.
That's no excuse, the transconductance based differential amplifiers like MAX4208 don't rely on resistor matching.
Probably, they can't produce instrumental amp due copyright issue.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20360
  • Country: gb
  • 0999
Re: Diffrential amp design
« Reply #14 on: January 24, 2020, 12:32:24 pm »
I looked for cheap instrumentation amplifiers a while ago and didn't find any. There are probably 10 different Chinese op amp manufacturers and you can find everything from 500MHz op-amps to low offset op-amps, but it seems no one has ever heard of instrumentation amplifiers. I think it's probably because trimming the resistors is expensive.
That's no excuse, the transconductance based differential amplifiers like MAX4208 don't rely on resistor matching.
The MAX4208 is very noisy compared to the AD620, around ten times as bad, so completely unsuitable for μV level signals.
https://datasheets.maximintegrated.com/en/ds/MAX4208-MAX4209.pdf
https://www.analog.com/media/en/technical-documentation/data-sheets/AD620.pdf
 

Online Marco

  • Super Contributor
  • ***
  • Posts: 7044
  • Country: nl
Re: Diffrential amp design
« Reply #15 on: January 24, 2020, 04:29:48 pm »
That's not really the point, noise is not inherent in the architecture, AD8129 has 4.5 nV/√Hz for instance ... the architecture can target any niche in-amps are targeted at. So if Chinese manufacturers want to build differential to single ended amplifiers without trimming, it's there.
 

Offline ali_asadzadehTopic starter

  • Super Contributor
  • ***
  • Posts: 1967
  • Country: ca
Re: Diffrential amp design
« Reply #16 on: January 25, 2020, 07:03:18 am »
I know that we should do something,if we do not do something, we are just limiting ourselves to the high tech giants to drink the juice out of our pockets with out dated parts at high costs!
ASiDesigner, Stands for Application specific intelligent devices
I'm a Digital Expert from 8-bits to 64-bits
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20360
  • Country: gb
  • 0999
Re: Diffrential amp design
« Reply #17 on: January 26, 2020, 11:34:29 am »
Probably, they can't produce instrumental amp due copyright issue.
Since when has that bothered Chinese companies?

That's not really the point, noise is not inherent in the architecture, AD8129 has 4.5 nV/√Hz for instance ... the architecture can target any niche in-amps are targeted at. So if Chinese manufacturers want to build differential to single ended amplifiers without trimming, it's there.
I admit, I'm not familiar with transconductance based differential amplifiers. Do they have high input current noise? The AD8129 has ten times the input current noise, of the AD620, at high frequencies: 1pA/√Hz vs 0.1pA/√Hz at 20kHz and much higher 1/f noise, so at 10Hz, so nearly 20pA/√Hz vs just over 0.2pA/√Hz. I know the AD8129 an RF part and not optimised for low frequency operation like the the AD620, so we're not comparing like with like.

Do you know of a part more suited to the original poster's application?
I know that we should do something,if we do not do something, we are just limiting ourselves to the high tech giants to drink the juice out of our pockets with out dated parts at high costs!
What do you think we need to do?

The semiconductor manufacturers are not ripping us off. Some things, such as low noise, low offset, high CMMR, high accuracy, DC amplifiers are technically difficult to achieve and are costly to make. New technology has made many things cheaper, such as achieving higher speeds, more complex designs on the same piece of silicon, but not everything has progressed at the same rate. Basic laws of physics can often get in the way, which gives us diminishing returns in certain areas such as low noise design.

I've already told you how you can nearly half the cost of your design: replace the AD620 on the right hand side of the drawing with a cheaper, ordinary op-amp.

EDIT: Corrected decimal point on resistor value, in schematic.
« Last Edit: January 26, 2020, 10:56:00 pm by Zero999 »
 
The following users thanked this post: nctnico, SiliconWizard

Offline 741

  • Frequent Contributor
  • **
  • Posts: 406
  • Country: gb
    • Circuit & PCB Design (small PCB quantities OK)
Re: Diffrential amp design
« Reply #18 on: January 26, 2020, 07:48:24 pm »
"INA333 Micro-Power (50μA), Zerø-Drift, Rail-to-Rail Out Instrumentation Amplifier"  :-+

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20360
  • Country: gb
  • 0999
Re: Diffrential amp design
« Reply #19 on: January 26, 2020, 08:10:32 pm »
"INA333 Micro-Power (50μA), Zerø-Drift, Rail-to-Rail Out Instrumentation Amplifier"  :-+
It's noisy though, compared to the AD620 at 50nV/√Hz, vs 9nV/√Hz, which I think could be an issue as they're dealing with µV level signals.
 

Offline MasterT

  • Frequent Contributor
  • **
  • !
  • Posts: 851
  • Country: ca
Re: Diffrential amp design
« Reply #20 on: January 26, 2020, 10:00:47 pm »
"INA333 Micro-Power (50μA), Zerø-Drift, Rail-to-Rail Out Instrumentation Amplifier"  :-+
It's noisy though, compared to the AD620 at 50nV/√Hz, vs 9nV/√Hz, which I think could be an issue as they're dealing with µV level signals.
What I see, ad620 noise goes up to 300 nV in the sub-Hz frequency. Application has filter at 0.4 Hz bandpass, so only this band matters.  I'd agree that AZ OPA is better option in this case, since noise flour stays flat for them. Logically, AZ circuitry just "null out" low frequency noise and becomes a winner.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20360
  • Country: gb
  • 0999
Re: Diffrential amp design
« Reply #21 on: January 26, 2020, 10:54:48 pm »
"INA333 Micro-Power (50μA), Zerø-Drift, Rail-to-Rail Out Instrumentation Amplifier"  :-+
It's noisy though, compared to the AD620 at 50nV/√Hz, vs 9nV/√Hz, which I think could be an issue as they're dealing with µV level signals.
What I see, ad620 noise goes up to 300 nV in the sub-Hz frequency. Application has filter at 0.4 Hz bandpass, so only this band matters.  I'd agree that AZ OPA is better option in this case, since noise flour stays flat for them. Logically, AZ circuitry just "null out" low frequency noise and becomes a winner.
That's what you might think, but the data sheets say otherwise.
https://www.analog.com/media/en/technical-documentation/data-sheets/AD620.pdf
http://www.ti.com/lit/ds/symlink/ina333.pdf

Input voltage noise from 0.1Hz to 10Hz is 1μVPP for the INA333 and 0.28μVPP for the AD620, at gains above 100, which the original poster's design has. The only advantage the INA333 has, noise wise is lower input current noise at 2pAPP vs 10pAPP, so might be better if the input is high impedance, but it's only rated to 5.5V single supply and the original posters design is +/-5V.
 

Offline GigaJoe

  • Frequent Contributor
  • **
  • Posts: 524
  • Country: ca
Re: Diffrential amp design
« Reply #22 on: January 27, 2020, 12:24:07 am »
I'm wonder ... if alternatives was observed ... as an example from the first page of AD620 PDF ?
like AD8221 ?
or replacement by regular 3-opamp schematics, with much wider choice for a regular opamp ...
 

Offline MasterT

  • Frequent Contributor
  • **
  • !
  • Posts: 851
  • Country: ca
Re: Diffrential amp design
« Reply #23 on: January 27, 2020, 12:30:32 am »
...
That's what you might think, but the data sheets say otherwise.
https://www.analog.com/media/en/technical-documentation/data-sheets/AD620.pdf
http://www.ti.com/lit/ds/symlink/ina333.pdf

Input voltage noise from 0.1Hz to 10Hz is 1μVPP for the INA333 and 0.28μVPP for the AD620, at gains above 100, which the original poster's design has. The only advantage the INA333 has, noise wise is lower input current noise at 2pAPP vs 10pAPP, so might be better if the input is high impedance, but it's only rated to 5.5V single supply and the original posters design is +/-5V.
The problem with DS is that there is no 0.01 - 1 Hz line, and looking into similar IC from TI TLV333 , where noise spec is split into 0.01 - 1 and 0.1 - 10 ranges, I see 0.3uV and 1.1uV, that's more than x3 times.
Right, it's just analytics, but if I go father into <0.5 Hz than difference with ad620 where line is sharply rising would be more obvious.  Theoretically, I don't know if China have any part numbers, very unlikely anyway.
 Regarding voltage, OP was going to redesign circuits and why keep same voltage?
 

Offline OwO

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: Diffrential amp design
« Reply #24 on: January 27, 2020, 02:50:20 am »
Oh I didn't realize it's using a gain > 100. In that case a discrete design with 3 opamps can achieve >100dB CMRR easily because the first stage provides 40dB of common mode rejection.
Email: OwOwOwOwO123@outlook.com
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf