Probing and scoping MegOhm circuits?

C

cupoftea

Advanced Member level 6
Advanced Member level 6
Joined
Jun 13, 2021
Messages
3,318
Helped
65
Reputation
132
Reaction score
153
Trophy points
63
Activity points
17,230
Hi,
We have circuits with multiple MegOhm resistors in them. Rails are mostly 3V3 and 5V.
Resistors are as high as 22MEGs in places.

If probing the attached with say a scope probe, would you recommend a 100:1 probe?

Also , please confirm that the "22MEG" circuit is useless?......i cant see any property of a photodiode that could ever help the collector of the NPN to ever go high?
Must admit that 22MEG resistor does look like an "elephant in the room" to me though?
(LTspice of "22MEG" circuit also attached)
 

Attachments

  • MEGOHM.png
    MEGOHM.png
    32.2 KB · Views: 44
  • 22MEG cct_.png
    22MEG cct_.png
    37.2 KB · Views: 38
  • 22meg cct.zip
    755 bytes · Views: 21
Last edited:

Sort by date Sort by votes
Bias for that transistor does not look right. What is the diode input,
pulse ? If so freq and amplitude....?

1742856580615.png
 

Upvote 0 Downvote
"22MEG" circuit is obviously an amplifier for AC light signals. It can work, hard to tell if component values are reasonable. Above about 5 uA, photo diode saturates and signal gain drops.
 

Upvote 0 Downvote
Circuit bias makes no sense to me. And output in sat for all photodiode currents.

1742904482374.png
And Zin at photodiode looks (dominated by 47K) :

 
Last edited:
Upvote 0 Downvote
Thanks, thats what i thought, how can that transistor ever be turned off by anything that the photodiode does?
 

Upvote 0 Downvote
Thanks
photodiode AC current
Click to expand...
Thanks, but all the photo diode datasheet that i have seen do not give an AC output current....its just DC current in proportion to the light exposure.

I wonder if you are referring to Johnson noise or something?, and how this will be more, the more is the DC current?...i think i read that in an application note somewhere......you detect photodiode current by realising that the higher the dc current, the higher is the Johnson noise in the high value R's through which this tiny current flows...and so you literally measure the amplitude of the noise so as to infer the magnitude of the DC current?
 

Upvote 0 Downvote
This looks fine for cutoff during sunrise or after a solar eclipse for a couple seconds due to 22M * 0.1uF.

But why do you show a circuit without a purpose or specs to ask a question?

The 22Meg has a very strong purpose as a one-shot and the collector R limit of 10 uA means the Rbe = 26/Ic = 2.5 Meg raises Vbe to about 10% of the 5V supply.

Any small glass diode or PD with 0.5uA/uW will detect daytime as long as the transistor has a current gain greater than the required current ratio from cutoff to saturation = 46 based on 22M / 0.47M.

The only part that does not do much is the 47 k resistor.

The biggest problem with this circuit is the derivative sensitivity is so high compared to the proportional, that it will generate a 1shot cutoff every time the sun appears suddenly after a cloud but not fog. (unless that was the intent of this cct)
 
Last edited:

Upvote 0 Downvote
Sorry for being unclear. See below what I meaned with photodiode AC current. Of course there's a DC bias superimposed, but the circuit can well detect 2 uA AC magnitude.

We don't know the actual circuit purpose, but it can work, as previously mentioned.

 

Upvote 0 Downvote
A little noise, I choose 1 uA RMS, due to high Z enable coupling, makes this circuit
a challenge. Timestep chosen 100 nS, lower would worsen results I would posit.



Of course we can sim this circuit to death, still not knowing what input is and application and......

Addendum, I see I missed C at E-B junction of transistor, adding that sim tolerated the 1 uA noise. My bad.

Note lowering or eliminating the 47K would add to the noise problem, so probably not a good idea. Also
beta of transistor probably a factor on noise immunity......

Interestingly I eliminated current noise source, used a 1V RMS noise source coupled by 5 pF to input, that
caused significant output noise problems.
 
Last edited:

Upvote 0 Downvote

If we assume the circuit was copied from a commercial design then the test pulses ought to be longer than the C1R3= 2.2 s time constant and with a slow rise time to mimic nature.

However if this detector is looking a pulse modulated IR light during sunlight with attenuation of fog attenuating both optical events then a higher frequency pulse can be blocked by a higher current pulse as the R2 22Meg gets shunted by the C1 coupled 1 MegR1. So the high frequency sensitivity is blocked by both too low a current pulse and too high a current pulse at too high a frequency limited by R1C1..

The purpose is getting clearer but still foggy.
 

Upvote 0 Downvote
Interestingly I eliminated current noise source, used a 1V RMS noise source coupled by 5 pF to input, that
caused significant output noise problems.
Click to expand...
Thanks, and to probe this cct is going to need a home brew probe probably coax,
and a lot more than 100:1...and very low probe capacitance.
Home brew only?
 

Upvote 0 Downvote
cupoftea said:
Thanks, and to probe this cct is going to need a home brew probe probably coax,
and a lot more than 100:1...and very low probe capacitance.
Home brew only?
Click to expand...
Knowing that the Vbe is less than 500 mV with a 10 Meg probe will draw less than 50 nA meanwhile the base current for Vbe= 500 mV is several uA. Do you really think 50nA shunt on the base will make any errors? This is because the incremental impedance is a lot lower than the linear equivalent = Vbe/Ib.

Below the variable 0 to 10 uA current source is series modulated with the left NPN and clock. while a 10 Meg probe is added to Vbe with current shown.

 
Last edited:

Upvote 0 Downvote
Post #1 assumption was "circuit is useless". Having shown that it can serve a purpose doesn't necessarily mean that it also has a useful application. The application range was never mentioned, unfortunately.

Returning to "how to probe". For high impedance nodes (photo diode, base circuit), you can't use standard probes. Best tool would be a high impedance (FET input) buffer. Question is, do you actually need to probe it in operation?
 
Upvote 0 Downvote
Frank, If you read my above comment I am sure you would agree, even a 1 Meg probe would not affect the readings when marginally saturated with 2 uA pulsed input with 50 nA or even 500 nA from 500 mV= Vbe even though it is only 485 mV with Ic = 12 uA.

Naturally if the input is 0 to 2 uA a FET buffered DIY or TEK probe is desirable but unlikely useful for a non-event since the function of the circuit can be tested with a pulsed >=2 uA input.

I agree with your comment on usefulness , as the 22 Meg is essential to make it work as a sunrise or rise in light current detector to divert the 22 Meg base current.
 
Last edited:
Upvote 0 Downvote
Attaching 1 Meg probe to base node kills bias point, also input capacitance affects AC band width. 10 Meg has only small effect and would be o.k.
Click to expand...
Thanks, i am not aware of too many 10Meg probes...the common ones i know of all have 1MEG shunt and 9 MEG series if 10:1, or 1MEG shunt and 99MEG series if 100:1?
 

Upvote 0 Downvote
You must log in or register to reply here.

Similar threads