Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Amp after RF detector not working as expected

Hawaslsh

Advanced Member level 4
Joined
Mar 13, 2015
Messages
102
Helped
5
Reputation
10
Reaction score
5
Trophy points
18
Location
Washington DC, USA
Activity points
1,279
Hello,

I am working with a simple RF diode detector circuit followed by a high gain op-amp with input offset voltage compensation.
Picture1.png
Above is a picture of the RF detection circuit alone. The circuit is capacitively coupled with a pair of diodes rectifying the input. A filter cap is used to create the DC voltage and the 1 Meg is there as a load. This works as expected, its pretty simple. Terminating the input with a 50 ohm load will cause the output voltage to drop close to 50 uV. If I apply a RF signal, the DCout voltage is proportional to the input power; for example at 11 GHZ, -20 dBm input power --> DCout = 40 mV.
Picture2.png
In reality I will be trying to detect much smaller powers than -20 dBm, and will need some pretty serious amplification to reach the 10 bit resolution of my 5V ADC. In order to do so, I added a non-inverting op-amp after the detection circuit. The topology came from an analog devices' application tutorial, and features an external offset voltage adjustment using a potentiometer. For my set of tests the feedback resistor was set to 10K Ideally giving me a gain of 11. I adjusted the pot to be as close to center (~0V) as possible.
Given the results from the detection circuit alone: I expected to see an output voltage of ~550uV with a 50 ohm termination on the input, and a ~440 mV output with a -20 dBm 11 GHz input. However, in reality, the op-amp sits at the positive rail. regardless of the RF input. If I probe the inputs to the amp directly, both sit at 549 mV, regardless of any input OR any adjustment on the pot. (As a note: The pot is working, I can measure the voltage change at the pot's pin 2.)
Clearly I am missing something pretty fundamental here. I dont know why the amp is behaving as it is. I am going to keep experimenting, however, i wanted to see if anyone had any suggestions to try or as to why?
Thanks in advance,
Sami
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
47,502
Helped
14,052
Reputation
28,359
Reaction score
12,711
Trophy points
1,393
Location
Bochum, Germany
Activity points
276,222
ADA4841 has typically 3 µA input bias current, may be more. The voltage drop at 1 Mohm resistor can exceed the compensation range.

Why do you use this fast bipolar amp instead of a general purpose OP?
 

Hawaslsh

Advanced Member level 4
Joined
Mar 13, 2015
Messages
102
Helped
5
Reputation
10
Reaction score
5
Trophy points
18
Location
Washington DC, USA
Activity points
1,279
Why do you use this fast bipolar amp instead of a general purpose OP?
No real reason other than we have a lot of them. However, I'll try an op07 on a breadboard after the detection ctk to contrast the results

ADA4841 has typically 3 µA input bias current, may be more. The voltage drop at 1 Mohm resistor can exceed the compensation range.
Can you explain what you mean about the compensation range? I only choose 1Mohm because the diode's datasheet suggested it as a load.
From my limited reading so far: should I have chosen a load resistor closer to the parallel combination of the feedback resistors?
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
47,502
Helped
14,052
Reputation
28,359
Reaction score
12,711
Trophy points
1,393
Location
Bochum, Germany
Activity points
276,222
You reported to have checked the operation of the compensation circuit, so you should know that the voltage range is +/- 3 mV. 3 mV is also the typical voltage drop across the 1 Mohm resistor. If a little bit higher, the compensation fails. Reduce R9 to e.g. 100k to increase the compensation range.
 

danadakk

Advanced Member level 1
Joined
Mar 26, 2018
Messages
425
Helped
83
Reputation
165
Reaction score
96
Trophy points
28
Activity points
2,154
The bias current sourced thru ~ 2K yields 6 mV of offset, so for low G's that
does not explain output pegged to rail. Sure sounds like fdbk R is not connected
or much higher value than you think and part is acting like a comparator.

But the fact both inputs at 549 mV says there is fdbk. Got me......unless part is bad.


Regards, Dana.
 

Attachments


Hi-Q

Junior Member level 2
Joined
Oct 2, 2011
Messages
23
Helped
0
Reputation
0
Reaction score
1
Trophy points
1,283
Activity points
1,596
Guys, 3µA through 1Megohm makes 3V.

The 550mV at the inverting input is obvious with supply voltages of +-6V and a gain setting of 11. Just the Output hanging on the positive supply divided by the feedback divider.
But why is the non-inverting input only at 549mV?

The input protection diodes should only kick in at about 1.5V (ok, maybe less here because of the low current they have to conduct but there are two diodes in series, so it should be quite a bit above the voltage at the inverting input.

How did you measure the 549mV at the inputs?
Could the measuring equipment have loaded the circuit node and limited the voltage just to 549mV?
Would be some coincidence, but once in a while such things happen...
 
  • Like
Reactions: FvM

Hi-Q

Junior Member level 2
Joined
Oct 2, 2011
Messages
23
Helped
0
Reputation
0
Reaction score
1
Trophy points
1,283
Activity points
1,596
If the OP would reverse the diode polarities, the OP bias current could flow through the diodes and even generate a nice little forward bias. The temperature dependency would actually compensate a bit for the temperature dependency of the diode detector ;-)
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
47,502
Helped
14,052
Reputation
28,359
Reaction score
12,711
Trophy points
1,393
Location
Bochum, Germany
Activity points
276,222
Possible, but not achieving obviously intended zero bias rectifier operation. Mentioned OP07 has still nA input current and generates respectively mV offset error. Better use low offset JFET or CMOS OP.
 

Toggle Sidebar

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top