Voltage Drop at Breadboard

[IanB]

anbudroid,
Did you measure the voltage directly at the DC:DC converter's input and output terminals on the board?

Don't the pictures attached to the first post show these measurements?

[rdl]

Don't use the rails on the breadboard. Run long wires for power directly to and from the DC-DC converter to the devices being powered. Connect them in the closest hole(s) to the part's power pins. Maybe even in the same hole if it's not too tight. If that fixes the problem then you might have a bad breadboard or the existing connections are bad. If it doesn't fix the problem, then something else must be wrong.

[tggzzz]

Isn't debugging solderless breadboards so much ... fun. And such a good way to learn electronics too.

[Zero999]

anbudroid,
Did you measure the voltage directly at the DC:DC converter's input and output terminals on the board?

Don't the pictures attached to the first post show these measurements?

Yes, you're right, I didn't look.

[sleemanj]

There is no magic, it's resistance, the question is where is the resistance.  If we look at your pictures between the DC/DC and the arduino are MANY connections into the breadboard, the DC/DC into two holes, then a pair of jumper wires from there to the rail (4 connections), then a pair of jumper across the rail break (another 4 connections), then another pair of jumpers from the rail down to the nano (another 4 connections), then the nano's connection down itself, another 2, that is 16 connections (8 + and  8 -) to between the DC/DC and the nano, I am not surprised to see even a couple of ohms lost there just in connections.

Chuck a decent sized cap across each device's power connection on the breadboard and that will help matters (if they need helping, if everything works, why worry).

[windsmurf]

Breadboards may be categorically unsuited to some types of circuits, but they're of incredible value to beginners especially. There's a TON of value in being able to quickly lash together a circuit and play around. It takes far longer to solder something together, and making changes are even harder still. I think you guys forget that breadboards are, above anything else, a learning tool — you guys are experienced and knowledgeable, but for those of us still figuring things out, breadboards are incredibly useful.

Just don't buy cheap!

What is a good breadboard kit to buy? Thx.

[rdl]

What is a good breadboard kit to buy? Thx.

https://www.eevblog.com/forum/chat/3m-breadboards-for-cheap/

Note that the first post also has links to three other threads about breadboards.

[Peabody]

Julian Ilett has a recent Youtube video that appears to show a source of good, cheap breadboards:



The Ebay.com link for this item appears to be this:

https://www.ebay.com/itm/253085438896

[tggzzz]

There is no magic, it's resistance, the question is where is the resistance. 

Indeed. But don't forget the inductance of a wire, 1nH/mm being a rule-of-thumb. Hence a 6" wire (as commonly used with solderless breadboards) adds ~150nH.

Chuck a decent sized cap across each device's power connection on the breadboard and that will help matters (if they need helping, if everything works, why worry).

And don't forget to calculate the LC resonant frequency and see if it is anywhere in the circuit's/components' frequency response (N.B. not signal's frequency response!).

[anbudroid]

I'm not convinced the breadboard is this issue in this case. Whenever anyone mentions breadboard, there's always this discussion which is valid, but is also a distraction.

I'm more suspicions of the DC:DC converter.

anbudroid,
Did you measure the voltage directly at the DC:DC converter's input and output terminals on the board?

Yes sir, my first picture shows the DC-DC Output voltage at the board level itself. am not think the problem is because of DD-DC . this problem persist even if I omit the DC-DC and plug the bench power supply directly to the breadboard.

    After reading all of the discussion and investing some times around the breadboard I finally convinced myself that breadboard are only worth to work with digital signals and non-precious analog circuits. if we want to test any project with confident, first thing is "say no to breadboard"

    because am not conscious about this kids of error previously when I work with switching applications . thinks create problem only when I need analog.

[tggzzz]

After reading all of the discussion and investing some times around the breadboard I finally convinced myself that breadboard are only worth to work with digital signals and non-precious analog circuits. if we want to test any project with confident, first thing is "say no to breadboard"

Digital waveforms are analogue signals, ones that are interpreted by the receiver as being digital signals. If the analogue waveforms don't meet the input specification, the receiver can misinterpret them.

The only digital waveforms you are likely to encounter are in photon-counting and femto-amp applications.

N.B. the frequency content of digital signals is determined only by the risetime; the  period is completely irrelevant. For a little theory and a practical demonstration, see https://entertaininghacks.wordpress.com/2018/05/08/digital-signal-integrity-and-bandwidth-signals-risetime-is-important-period-is-irrelevant/

Modern jellybean logic has frequency content above 1GHz. Work out the current required to change the voltage of a 5pF capacitor by 5V in 1ns (typical input/voltage/time). Now realise that a 2cm wire on a solderless breadboard has 20uH inductance. Finally work out the induced voltage v=L^di/_dt. And then you won't use solderless breadboards for digital signals

Read Bogotin's rules of thumb https://www.edn.com/collections/4435129/4/Bogatin-s-Rules-of-Thumb

[Ian.M]

Digital logic on a breadboard is  not *THAT* bad - as long as you don't try to use faster logic than LS TTL and HC CMOS, and put the decoupling caps directly over or next to the chips, with their leads as short as possible into the holes closest to the Vcc and Gnd pins.

Clock inputs can be a problem, but the typical CMOS input protection diodes clamp overshot fairly well, and if you control slew rate with source termination you can usually eliminate glitches on the active edge that would otherwise cause double clocking.  Occasionally you may need to resort to source terminated twisted pair for a longer signal run, possibly with AC termination using a RC network at the receiving end to damp transients on the edges (100R + 22pF to 100pF).  Data inputs aren't usually a problem, as long as the clock rate is low enough for ringing to settle out before the next active clock edge.

It also helps if you are using a MCU with all the high speed stuff on one die, rather than a CPU with ROM, RAM and I/O spread over several breadboards.

[tggzzz]

Digital logic on a breadboard is  not *THAT* bad - as long as you don't try to use faster logic than LS TTL and HC CMOS, and put the decoupling caps directly over or next to the chips, with their leads as short as possible into the holes closest to the Vcc and Gnd pins.

I did specify modern jellybean logic, not mid-70s 40-year old families!

Lead inductance in DIL ICs could be a serious problem in the mid 80s.

Modern large devices are so susceptible they come with IBIS LCR specifications for each pin (including multiple Cs to different rails), plus the waveform for each different drive level.

[anbudroid]

[/quote]

Digital waveforms are analogue signals, ones that are interpreted by the receiver as being digital signals. If the analogue waveforms don't meet the input specification, the receiver can misinterpret them.


[/quote]


 you are right sir , but beginners  are not play more than 10Mhz clock signals with breadboard. and most of the Arduino players are much below this level.

[radiolistener]

I catch the same issue at breadboard which I bought on aliexpress. In my case, the root of cause was breadboard resistance. It leads to voltage drop down when circuit consumes a lot of current. In my case I was used high speed ADC with power consumption 300 mA. And voltage drop was about 0.3-0.6 V.

[tggzzz]

Digital waveforms are analogue signals, ones that are interpreted by the receiver as being digital signals. If the analogue waveforms don't meet the input specification, the receiver can misinterpret them.

 you are right sir , but beginners  are not play more than 10Mhz clock signals with breadboard. and most of the Arduino players are much below this level.

Sigh. That misapprehension never goes away.

My step generator has ~1MHz period, and has significant energy content above 1GHz. The period is completely irrelevant; the only thing that matters is the transition time.

If you want an anthropromorphic handwaving explantion, consider that when the signal has one edge, it doesn't "know" when the next edge is coming - it might be in 1ns, 1s, or never.

For a little theory and some practical demonstrations, see https://entertaininghacks.wordpress.com/2018/05/08/digital-signal-integrity-and-bandwidth-signals-risetime-is-important-period-is-irrelevant/

[rdl]

Out of curiosity, tonight I wired up a little test. Using one of my 3M No. 318 boards I input about 5 volts at the left end of the upper rails, ran it across to the right side and down the vertical rails, then back to the left end of the lower rails, a distance of nearly 17" (about 430mm). Actually double that for the entire loop. This meant it also went through 4 jumper wire connections (22 gauge tinned solid copper wire).

With no load, there was virtually zero voltage drop, maybe 1 mV at most. With two 47 ohm resistors in parallel as load (210 mA measured) I got about a 40 mV drop. This was from measuring directly at the input to the breadboard vs. measuring directly across the resistors.

Just thought people might find that interesting. If I get a chance, I may send a square wave through and see what it looks like on the scope.

[tggzzz]

Out of curiosity, tonight I wired up a little test. Using one of my 3M No. 318 boards I input about 5 volts at the left end of the upper rails, ran it across to the right side and down the vertical rails, then back to the left end of the lower rails, a distance of nearly 17" (about 430mm). Actually double that for the entire loop. This meant it also went through 4 jumper wire connections (22 gauge tinned solid copper wire).

With no load, there was virtually zero voltage drop, maybe 1 mV at most. With two 47 ohm resistors in parallel as load (210 mA measured) I got about a 40 mV drop. This was from measuring directly at the input to the breadboard vs. measuring directly across the resistors.

Then try knocking the breadboard and generally moving it in ways that you would do when experimenting with a circuit. See if anything changes.

Wait a couple od days, and repeat.

Just thought people might find that interesting. If I get a chance, I may send a square wave through and see what it looks like on the scope.

Measurement technique is important, so document the source characteristics, the scope, and how you connect those to the UUT.

[rdl]

A couple of pictures to look at. The square wave from my function generator is pretty sad, but for comparison purposes it'll have to do for now.

10MHz square wave at entry to board (blue) and at 47ohm resistor load 17" away (yellow)



20Khz sine wave (same situation, blue at entry, yellow at load)



I'll add this in case my previous description wasn't clear.



The voltages measured previously were at A and B across the rails. The signals are compared between points A and B. The resistor (47 ohm) was connected to the negative rail, not disconnected as in the picture. Two resistors in parallel were used for the voltage measurements.

[tggzzz]

14ns risetime <=> 25MHz bandwidth, so your scope isn't the limiting factor.

I suggest you look at the output of a logic signal. For bonus points, insert two 6" breadboarding leads in between the scope/signal and scope/ground. The 15pF scope probe tip capacitance is similar to a couple of CMOS/TTL loads.

[rdl]

I was not sure what more information another 6 inches of wire would add, after all, there is already ~425mm of conductor including 8 wire to board insertions, but I looked at it anyway.






I don't actually see much difference.

While I was at it, I measured the resistance of one 425mm length of breadboard rails (plus wire jumpers) at 18 ohms. If I did the math right, this agrees closely with the 40 mV voltage drop I recorded previously. The capacitance across the pair of rails (all 425mm) was measured as about 85 pF.

[bd139]

A point to bear in mind when using or not using breadboards: Sometimes signal integrity and layout doesn't actually matter that much!

I'm doing some sub 1KHz stuff with opamps at the moment which requires a lot of trial and error. I'm doing it on a 3M breadboard and am giving zero fucks

[rdl]

One more thing, and this is something that I've previously seen evidence of potentially being a real problem. This image is where the crosstalk is looked at on the long 425mm path. The blue trace is the actual signal on the "positive" rail (red jumpers), and the yellow is the induced signal on the "negative" (black jumpers) rail.



Here the same signal is running through one of the five hole vertical strips (blue trace) with the crosstalk that appears on the adjacent five hole socket strip (yellow trace). Doesn't look that bad actually.



I'd like to point out to anyone reading this thread that running a 10MHz signal through 425mm of conductor, including multiple wire to board insertion connections, is somewhere beyond "worst case scenario". I'm sure decent quality breadboards have a limit to their usefulness, but I don't know where it is and I'm not inclined to go looking today.

[bd139]

On that basis you could have some fun and use the power rails as a directional coupler

[rdl]

Here are the waveforms with and without the extra 200mm jump wire inserted after going through the full 425mm rail length. I previously said I didn't see much difference. Taking a second look, the one with the extra jumper is actually shifted slightly more. While definitely something to keep in mind, it's not likely to be a problem in any circuit I would build on a breadboard. Signal connections there should always be as short and direct as possible, not two feet long with multiple jumper wire breaks.