What if 10nF is cut in this circuit?

[T3sl4co1l]

What would be the difference between the 3 capacitors vs a single one of 10.11 MicroFarad?

Basically you get an impedance more like this,



rather than a single valley.

If the peaks inbetween are well controlled (by judicious choice of capacitor ESR), the result can be better (lower impedance over a wider frequency range).  If not (just stacking ceramics in parallel), the peaks can be quite high (several ohms), and that may have consequences for device performance.

The difference between cargo cult and science is modeling and measurement.

Tim

[Siwastaja]

See this thread: https://www.eevblog.com/forum/projects/decoupling-caps-value/?all
specifically replies 23, 27 and 30

[bonyz]

Before when I tried to short the right side or use resistor of 0 Ohm (isn't it 0 Ohm is same sa short?), the 0603  C25 capacitor at middle got accidentally desoldered from the solder heat. I transferred the original C25 capacitor to between the left capacitors  (shown in blue arrow from middle to left between C27 and C26 in original diagram, the controversial 3 parallel capacitors illustration you saw earlier) 

This occured because I had difficulty removing the remaining solders at the middle and I couldn't reinsert the C25 capacitor at original position. Also I'm not sure the C25 is still viable with all the heat hence this thread.

Anyway. What new noises were introduced in the above solders and heating and how they occur?

Also I know that at high frequencies. Impedances of capacitors got lowered and acting like short circuit that produces voltages of zero. In this case. It's decoupling capacitors. For value of 10uF. What kind of power line noises get filtered and at what frequencies? How do you compute for it?

[bonyz]

Here are clearer photos described in last message above.

Original area (before alternation made).



Following is alteration (C16 removed, R18 shorted to make it 0 Ohm, the procedure damaged C25 and I moved it between C27 and C26):



The solder heat has deformed the area. Could it have destroyed the capacitors?  Solder used.



When soldering 0603 components. What kinds of solder are you supposed to do? And how do you desolder or flatten the rough solder splattered surface so capacitors or components can be replaced?

[T3sl4co1l]

Use a bigger tip and preheat the board to 100-150C. Use paste flux.

Tim

[xrunner]

And how do you desolder or flatten the rough solder splattered surface so capacitors or components can be replaced?

Just removed and replaced some 0603 stuff recently. Add additional flux. Add more solder and use a chisel tip. Move the tip back and forth and the solder on both sides of the component will melt. Swipe off the component. Use desoldering braid to remove all the extra solder. Clean the area. Re-solder with new components (add more flux).

Look at some of the very good SMD soldering videos here -

https://www.youtube.com/@Androkavo

[themadhippy]

For value of 10uF. What kind of power line noises get filtered and at what frequencies? How do you compute for it?

1 / (2 × π × f × C) were f is your frequency in hertz and c is your capacitance in farads,that'll roughly tell you the equivalent resistance of the cap at that frequency.

[fourfathom]

You can sometimes find values in the datasheet for ESR and Self-Resonant Frequency.  Lead/trace inductance will also affect the SRF.  Of course it's even more complicated than the image above shows.  Here's more: https://www.johansontechnology.com/srf-prf-for-rf-capacitors

To unsolder small SMT parts I sometimes use two fine-tip soldering irons.  Better though is a hot-air rework tool with a small nozzle and a pair of tweezers.  Otherwise try the "back and forth heating large blobs of solder" method.

[rstofer]

Dave did a video:


That is some pretty poor soldering.  Use more paste flux to help the solder wet the surfaces.

You might try dabbing some solder paste (with flux) on the pads then placing the components.  While holding the component in place with curved tweezers, hit the joint with the soldering tip.

Wire solder works but it usually takes a dab of flux.  In fact, put the flux down first and place the component in the puddle.  Put a dab of solder on the iron tip and hit the joint.  There's no need to use the wire solder directly on the component.

[radiolistener]

Such capacitors are added to block RF noise at high frequency. Because high capacitance capacitors cannot block it due to high parasitic inductance.

If you remove it it leads to a higher RF noise on power line. It may lead to unstable work of your circuit, or it can still work but will be more sensitive to RF environment and interference from nearby electronics, it depends on exact case.

[bonyz]

Such capacitors are added to block RF noise at high frequency. Because high capacitance capacitors cannot block it due to high parasitic inductance.

If you remove it it leads to a higher RF noise on power line. It may lead to unstable work of your circuit, or it can still work but will be more sensitive to RF environment and interference from nearby electronics, it depends on exact case.

What if the circuit is totally battery powered. Can RF from radio stations affect it if the 3 capacitors were removed?  You mean radio stations or cellular phones can couple to the circuit pcb which acts like antenna?

[fourfathom]

What if the circuit is totally battery powered. Can RF from radio stations affect it if the 3 capacitors were removed?  You mean radio stations or cellular phones can couple to the circuit pcb which acts like antenna?

Yes, without a doubt.  I can tell you horror stories about fixing susceptibility to RFI in battery-powered devices.

[bonyz]

What if the circuit is totally battery powered. Can RF from radio stations affect it if the 3 capacitors were removed?  You mean radio stations or cellular phones can couple to the circuit pcb which acts like antenna?

Yes, without a doubt.  I can tell you horror stories about fixing susceptibility to RFI in battery-powered devices.

Are their faraday cage like enclosure for circuit? how does it look like? I imagine like a screen cage? Haven't seen one yet.

Also if your op amp has bandwidth of only 2kHz for certain gain (like 100). Can RF noise of higher frequencies still affect it? How? maybe resonating the op amp internal parts?

[wraper]

Here are clearer photos described in last message above.

Original area (before alternation made).

With such poor PCB layout adding smaller caps would not make any improvement even if it actually mattered.

[WimWalther]

Back in the Olden Times, capacitors usually had wire leads, and thus significant lead inductance and low self-resonant frequencies.

Might be worth mentioning that leaded, spiral-wound caps are still the norm in point-to-point builds and circuits with high B+ voltages.. like tube audio gear. The well-regarded (...) Sprague "orange drop" parts being one nearly ubiquitous example, or the Illinois MPE / MPW series.

[wraper]

BTW this schematic is "genius" in many ways. Not only they decouple opamp power with 2 capacitors. They put 3 capacitors to the output of an opamp, two of them to GND (one 47uF tantallum  , I wonder if they added this much capacitance because this garbage was oscillating), another to VCCA when there should be none at all . Then there is another 2 cap nonsense with C10, C12 connected to the same poor output of an opamp already abused by 3 another capacitors. Then they use 2 caps for AP431 instead of one for no good reason.

It is part of the Arduino (see image below)...



The input can either be 3.3V or 5V. What is the purpose of the inductor and zener? If the 5V is pure 5V DC, then no purpose for them? Or could the purpose be to clean the 5V DC that could be somehow messed up by the Arduino crystal?

[bonyz]

Are there any generic circuit to create VREF that has fewest components and time tested that you can use as voltage reference to use in any instrumentation amps and op amps?

Also what particular noises are introduced if you put all of them in breadboard, or buy separate modules (with separate functions) and connect them with a 6 inches pair of wire? What size and length of wire that can produce less distortions and what inductive and capacitive interferences that can be created in such short wires?

[wraper]

It's not that there are just too many components. Those capacitors do harm, not good. OPAMPs are not supposed to drive highly captative load (unless specially designed to do so), they misbehave when doing so. And adding capacitance from VREF to VCCA literally worsens performance as noise from VCCA can now freely pass into VREF.

Are there any generic circuit to create VREF that has fewest components and time tested that you can use as voltage reference to use in any instrumentation amps and op amps?

The fewest components would be to use internal MCU VREF. Also it's not clear why 1.5V is needed. If 2.5V is OK, you could get it straight from AP431, which BTW is not that good of a reference, just cheap.

[T3sl4co1l]

If this is a for an ATmega family Arduino, external ref is desirable; the internal ref on those is rather a joke.  XMEGA and up are passable though (more bits ADC too).

There are '431 variants (TLV431, LMV431, ZR431L, etc...) with lower threshold voltage, also lower current consumption and lower noise are possible.  '431 (in the tighter precision grades) is adequate for the basic ATmega ADC (10 bits, with hardly 8 of them guaranteed).

Note that ADC performance is degraded below some minimum recommended Vref.  Better to amplify small signals, than to reduce Vref.  (Some gain is also possible, on some of these: XMEGA ADC has an iterated gain-of-2 feature for example.)

Tim

[T3sl4co1l]

Back in the Olden Times, capacitors usually had wire leads, and thus significant lead inductance and low self-resonant frequencies.

Might be worth mentioning that leaded, spiral-wound caps are still the norm in point-to-point builds and circuits with high B+ voltages.. like tube audio gear. The well-regarded (...) Sprague "orange drop" parts being one nearly ubiquitous example, or the Illinois MPE / MPW series.

Non sequitur -- almost all film caps are rolled, regardless of ESL.  The most common differentiator is whether there's a connector tab (typical of canned oil types) or schoopage (everything else).

In fact Orange Ddrops have quite good AC performance, but for the steel leads (see 716P instead), and the high cost (TDK, KEMET and others have cheaper parts of comparable ratings).

Tim

[wraper]

Back in the Olden Times, capacitors usually had wire leads, and thus significant lead inductance and low self-resonant frequencies.

Might be worth mentioning that leaded, spiral-wound caps are still the norm in point-to-point builds and circuits with high B+ voltages.. like tube audio gear. The well-regarded (...) Sprague "orange drop" parts being one nearly ubiquitous example, or the Illinois MPE / MPW series.

Non sequitur -- almost all film caps are rolled, regardless of ESL.  The most common differentiator is whether there's a connector tab (typical of canned oil types) or schoopage (everything else).

In fact Orange Ddrops have quite good AC performance, but for the steel leads (see 716P instead), and the high cost (TDK, KEMET and others have cheaper parts of comparable ratings).

Tim

I do not agree with that, many (especially small) film capacitors in plastic epoxy filled box are stacked, not rolled. Does not have much practical difference though as long as terminals are attached to metal sprayed ends of the stack/roll.

[bonyz]

The VCCA and VREFs and all those circuit are powering up this EKG/EMG circuit:



complete schematic:



Now do all those capacitors justify it because it is powering the EKG/EMG?

I didn't share it before because people got nervous about medical electronics and afraid the newbie may use it at hospital. But don't worry. I follow the warning note by Olimex by heart which is:

"This evaluation board/kit is intended for use for engineering development, demonstration, or evaluation purposes
only and is not considered by OLIMEX to be a finished end-product fit for general consumer use. Persons
handling the product must have electronics training and observe good engineering practice standards. As such,
the goods being provided are not intended to be complete in terms of required design-, marketing-, and/or
manufacturing-related protective considerations, including product safety and environmental measures typically
found in end products that incorporate such semiconductor components or circuit boards."

Where are you supposed to add the Isolator? It's connected to laptop usb and battery powered only anyway. Do you know of other schematic like it but medical grade? I'm just curious how it differs to $10,000 EMG. What chips are used in the latter?

About the Electrode:



"**The cable features three electrodes – two data electrodes (1 channel) and DLR electrode
(feedback). If you use more than one SHIELD-EKG-EMG you can use cables without DLR
for every shield after the first."

So VREF is connected to the right foot as part of DLR feedback. Are all the VCCA, VREF design justified? Or do you still consider them poor design?

Again I don't use it as medical device so don't worry. Heck I don't even use it as ECG/EMG because after you try it once. You are tired of it. I'm just curious how it compares to $10,000 version. Is there such that uses superconductor as detector element?

[wraper]

The VCCA and VREFs and all those circuit are powering up this EKG/EMG circuit:

I see that V_REF actually a virtual ground, not voltage reference. Still it's a poor design which can only be explained by a poor knowledge of whoever designed that.
Input ESD protection by PGS1/2 provides no actual protection. Why in hell anodes of D1/D2 pretection are connected to virtual ground and not actual ground? IC2B supposedly adjusts DC output voltage of in-amp equal to virtual ground, but what it's supposed to achieve compared to directly connecting REF input to virtual ground directly when there is C7 DC block anyway and offset voltage is pretty much irrelevant? (EDIT, Looked into INA321 datasheet and it looks like it's a result of a bastardization of ECG circuit found there) Neither I'm convinced it needs DC block capacitors at each amplification stage and as a side effect requiring DC bias resistors for inputs. Using X7R cap for active filter is questionable. IMHO you could easily throw out 50% of the parts from this circuit and make it work better at the same time.

[T3sl4co1l]

I see that V_REF actually a virtual ground, not voltage reference. Still it's a poor design which can only be explained by a poor knowledge of whoever designed that.

How so?  It's buffered (op-amp output), not virtual ground (op-amp input held at GND by action of feedback through a network).

If the op-amp buffer were handled properly (damping resistor feeding bypass caps), and the capacitance increased a bit (I'd use one or two 1uF+ ceramic placed very near D1 and D2).

Also if the input signal is very low, D1 and D2 could just be back-to-back tied with VREF.  Instead of clamping positive to VCCA (which needs a local 1uF+ for the same reason).

Input ESD protection by PGS1/2 provides no actual protection. Why in hell anodes of D1/D2 pretection are connected to virtual ground and not actual ground? IC2B supposedly adjusts DC output voltage of in-amp equal to virtual ground, but what it's supposed to achieve compared to directly connecting REF input to virtual ground directly when there is C7 DC block anyway and offset voltage is pretty much irrelevant? (EDIT, Looked into INA321 datasheet and it looks like it's a result of a bastardization of ECG circuit found there) Neither I'm convinced it needs DC block capacitors at each amplification stage and as a side effect requiring DC bias resistors for inputs. Using X7R cap for active filter is questionable. IMHO you could easily throw out 50% of the parts from this circuit and make it work better at the same time.

This.

Tim

[CatalinaWOW]

Tim comment and question about why the "standard' values for pullups and bypasses probably has a self fulfilling economic component.  As production volumes of these components rise the cost goes down.  Since a wide range of values work, pick the cheapest one.

Paralleled filter capacitors largely come as Tim mentioned.  Many beginners for time or capability reasons don't fully analyze every detail of their circuits.  And once upon a time the app notes for many common linear ICs recommended exactly this configuration.  The supply is often the least interesting part of the design so the app notes was copied and pasted in.  Thus creating a tradition which is difficult to fight.