Keyboard Shortcuts
Likes
Search
Need help to design a transimpedance amplifier
¿ªÔÆÌåÓýHi there, Please, Please see an Idea I have posted in the Temp Folders of this group. (Photo Amplifier15.asc) Best regards, Michael P Kiwanuka To: LTspice@... From: LTspice@... Date: Fri, 5 Aug 2016 09:10:31 +0200 Subject: Re: [LTspice] Re: Need help to design a transimpedance amplifier ? I tried to answer a couple of days ago, but it didn't come through, so I'll write it again: The noise you observe compes from the fact that you have these big 100k resistors in the non-inverting inputs. The OPA2846 has a very high NOISE CURRENT; it is designed for low-impedance sources. There should be no resistor in the non-inverting inputs, or very low (similar to the APD's equivalent resistance); if used, this resistor should be bypassed by a capacitor, for near-zero impedance within the circuit's BW. That's VLN 101. Le 04/08/2016 ¨¤ 09:31,
t.obulesu@... [LTspice] a ¨¦crit?:
?
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
I read through a few post, you want to detect from 200nA to 100uA with BW>=1MHz. You can do it in one single stage with feedback resistor of 50K. The output range is from 10mV to 5V. This is not hard to do. I uploaded the circuit and simulation using capacitance of APD of 5pF to show you get 1MHz BW.
I wrote my own excel program to calculate total noise. The numbers are for LTC6268, but if you go to the bottom, I have numbers for quite a few opamps you can plug into for noise simulation. I have short explanation how I calculate noise. I am not going to go through that in detail as this is LTSpice forum. Google and you'll find many article how to calculate noise of TIA. You can compare with my short explanation in the excel file. read through the instruction if you care to use the program for other opamps. Good Luck. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¿ªÔÆÌåÓýWould you like to look at the Idea i posted in the Temp file(photo amplifier 15A) and somebody else reproduced it with different component values(photo amp) i think you may have to rethink your TIA noise strategy if you understand how it works. ?Please note the following correction:= ?[(wc)**2 = 1/(R1RL)]. ?Please attempt to show that the expression I have just quoted is the notch condition for the op. amp. n/w. comments are most welcome Michael To: LTspice@... From: LTspice@... Date: Thu, 4 Aug 2016 11:03:19 +0200 Subject: Re: [LTspice] Re: Need help to design a transimpedance amplifier ? Check the datasheet for the OPA 2846: Vn=1.2nV/sqrtHz, In=2.8pA/sqrtHz that means the OSI (optimum source impedance) is 428 ohms. The 100k resistors in the non-inverting inputs alone introduce 88uV (@100kHz BW) What is the transducer's impedance? Le 04/08/2016 ¨¤ 09:31,
t.obulesu@... [LTspice] a ¨¦crit?:
?
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
How come it took two days before my message were posted?
You look for an opamp, it's not the bias current you worry because you are only measuring 200nA lowest. Any opamp with bias current less than 10nA is plenty good. The most important thing you look for is input capacitance. Most of the opamps have over 5pF of input capacitance and that will increase the noise gain. You want to understand noise gain, read the explanation in my Excel file. I explain each and every component of the noise source and the relevant frequency range. I chose LTC6268 because the input capacitance is about 0.5pF. For your requirement, some BJT input opamp might be good enough, just look for low input capacitance. That said, you always see noise coupling into TIA circuits. I never saw a TIA circuit that dose not pick up all sort of noise.? All the noise calculation and the excel file I provided are ONLY for ideal theoretical noise limit. If you leave the bare circuit, you will see all the noise from around the surrounding. FYI, our contractor designed another TIA that has only a 1M feedback resistor and it pick up all sort of noise. We have to re-layout the board to put in the shield. A ground plane is an absolute MUST. The other reason is people have long trace at the -ve input. That will pickup noise every time. Layout is everything. This is beyond the scope of this forum. Circuit is always simple, it's the layout and shielding. You need to look at your layout, put in the shielding, then talk more about it. LTSpice is NOT the place to talk about this, go to some electronic forum and ask question. I feel I already way out of bound here already. You are only working with 50K feedback resistors. My circuit use 250M, and I designed with up to 2G resistors TIA circuits. Layout and shielding become the number one design. If you read my other post, 50K resistor will give you output range of 10mV to 5V( fit your 5V supply requirement). This is very easy signal to deal with. We deal with a lot lower signal successfully. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Hello to All,
This topic has a rather lengthy response indeed. I don't have a specific response, just a suggestion. I have a copy of a book that I have found very helpful in some of the designs I have don., You might consider getting a copy. Title: Photodiode Amplifiers Author: Jerald Graeme Publisher: McGraw Hill ISBN #: 0-07-024247-X It covers just about all aspects you will need in designing a good TIA. Good Luck with your design Pete |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Hello wms,
Just a tip for a workaround regarding the delay. Write the message in a text editor. Then open the message you want to reply with your browser(Chrome, Firefox,..). Click on REPLY. Immediately copy your message their and then immediately click on SEND. This seems to work. It looks like if one edit the message online (too long time), the message will takes day or will never arrive. Best regards, Helmut |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¿ªÔÆÌåÓýHi WMS, The text on your schematic has an error in the expression for the notch condition. ?The correct expression should be? (wC)**2 = 1/(R1RL). ?A quick dimensional analysis agrees with that expression. ?I invited readers to satisfy themselves that the expression is true for the network with the op. amp. ?A hint; ?assume that each capacitor has a value C and that R1>>RL1 where RL1 is the resistance of the inductor L1. WMS, how did you arrive at the components values you used in your schematic ? ?what analytical rules or principles did you use. Regards, Michael P Kiwanuka To: LTspice@... From: LTspice@... Date: Mon, 8 Aug 2016 13:42:09 -0700 Subject: [LTspice] Re: Need help to design a transimpedance amplifier ?
ok,,,,3 days went by and yahoo didn't post me.? Here it is again...I uploaded "PHOTOAMP.zip" to the .TEMP files.? WW
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
It's been 3 days since I posted and still have not shown up.
Regarding to noise pickup like from power supplies, power lines, tv etc. It's very common, any reasonable high transimpedance gain circuit need to have shield to cover the circuit. This kind of noise has nothing to do with the normal noise calculation like voltage and current noise. From my experience, other than the very high speed TIA that has feedback resistor under 1K, all the TIA circuit need to have shield. we just had to relayout a TIA circuit the contractor design that is only 1M feedback resistor. It was picking up noise big time. For minimum of 200nA signal, that is huge signal, you don't even come close to the voltage and current noise issue. With a 50K feedback resistor, you have 10mV minimum signal. This is big signal!!! Layout is critical, the -ve of the opamp has to be point to point with minimum trace length. Most people I've seen have long trace on the -ve input, that is inviting noise pick up. All these has nothing to do with LTSpice, it's good common sense EE. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Hello All,
Thanks for your response... Well I have tested almost all inputs in LTSpice as per my design criteria:...it works fine on the screen...
??? ??=(V? (???)? ?? (???) ) / ???(???) ???? =4.5?? 0.1?5 ¦ÌA=880 ?¦¸ ??? ??¡Ü ??? ?? ?? ??¡Ü 0.18 pF at 1 MHz pole ???? ¡Ý (??+?? ) / (?? ?? ???)??? { ??= total capacitance of the photodiode and op-amp (0.5pF + ~ 5pF); } ¡Ý (? ??+?.???+?.?? ??) / (?????? ???.??^2) ?¡Ý45 ???
Please correct me if am wrong.... |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
try use the excel simulation file I uploaded a while back and enter in the numbers of your other opamps from the datasheet. If you read through the explanations, they will show you how to enter the numbers into the ones in red color. You can find the numbers from the datasheet. Then you adjust the feedback capacitor Cf+Crf on line 31 to get the 1MHz BW and look at line 47 to 51 for frequency response and inband noise. The total noise is the rms noise. If you shield the circuit properly, you should see the result similar to the number. Remember this is rms noise.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
John Woodgate
¿ªÔÆÌåÓýRE: Layout is critical, the -ve of the opamp has to be point to point with minimum trace length. Most people I've seen have long trace on the -ve input, that is inviting noise pick up. All these has nothing to do with LTspice, it's good common sense EE. ? This is a *very important general point*, not just for TIAs. ?The minus input of an op-amp has a very low impedance, due to the negative feedback, so it might appear that a long trace is OK, BUT it also has 'infinite' sensitivity, ?Keep the trace as short as possible. If you can't, put a 1 k resistor in series, close up to the op-amp pin. But don't do that if you are using the input as a current sink for an inductive pick-up, so as to get a flat frequency response. ? With best wishes DESIGN IT IN! OOO ¨C Own Opinions Only J M Woodgate and Associates Rayleigh England ? Sylvae in aeternum manent. ? From: LTspice@... [mailto:LTspice@...]
Sent: Wednesday, August 10, 2016 2:54 AM To: LTspice@... Subject: [LTspice] Re: Need help to design a transimpedance amplifier ? ? It's been 3 days since I posted and still have not shown up. |