what do they do in this video?

Status
Not open for further replies.
A

AdmiralCrunch

Well-known member
Hi

I found this video here: https://www.youtube.com/watch?time_continue=878&v=FqQDDpIgNyo
of how their development-process looks like.. nice and interesting.. but there is one thing i dont't understand:

from 12:48 on, he says now it's about testing the calibration.. therefor they have some small pots on the PCBs ..

my question is, what would be a use case for this? what would one need to calibrate? .. i mean, if it works, it works.. can't imagine what would one need to calibrate if the schematicss is fine(?)
 
looks like its used to stabilize the circuit’s signals, alot of electronics have these pots, even televisions. once adjusted, theres no reason to tweak it again unless someone plays with them or the daughter board interface was replaced
 
tonton81 said:
looks like its used to stabilize the circuit’s signals
Click to expand...

but how come? .. I mean, when I do the schematics .. then when soldering and wirering.. äi thought I have to take care, that everything is stable(?) .. I don't get it why there would be a need to adjust it.

also I don't understand how this adjustment works with a pot.. I know how a pot works.. it gves me a dynamic resistance.. but where would one position such a adjustment-pot in a circuit?
 
I didn't watch the video - but parts are mass produced with some tolerance.

Some things like pots can be fine tuned in software for their real 'observed' values to get the expected range - it can also be done as noted on TV's where there is a critical range the hardware had to be adjusted for functionality outside of software.
 
Parts tolerence is the big reason - Common parts are 5% but 10% or worse parts show up when you get into exotic sizes or low costs. For digital pullups 10% does not matter but for things that have analog processing in some form successive tolerance variation, especially in gain of transistor driven components means you need a twiddle pot to standardize things into your device. If you have a look at cheaper op amps you will normally find pretty broad specs that your design will need to handle.

See also:

https://en.wikipedia.org/wiki/Analog_computer for devices that were nothing but twiddle pots and thankfully no longer a thing.

Digital processing means that in many cases you just ADC your signal and correct things in digital form but in some cases you still need to tweak an input circuit or equalise an output.

In the RF domain you can also have correction needed for run to run variation in things like PCB thickness, moisture content of the board and/or conformal coating and component temperature induced gain or dimension changes.

And finally you can have 'the math was hard so we just put a pot there instead of a restive divider' where the pot is not a unit to unit adjustment so much as a low volume/single unit production run design cost saving measure.
 
did i understand it correctly?

many components in a circuit do many things, which means they consume current. the more components there are, the more does the current vary .. so one corrects this at some ponts in the current with this pots giving resistance?
 
That's not forcibly the point. Ideal components do not exist in reality. Each capacitor is two resistors and one inductor at the same time, each inductor is two resistors and one capacitor at the same time, for example. Each copper trace on a PCB is a resistor, a capacitor and an inductor at the same time. And all this with production tolerances of up to +/-20%. That's the reality. Since not all side effects of these tolerances and parasitic resistances/capacitances/inductances can be predicted and be taken into account, circuit designers add variable components like trimmer potentiometers or capacitors to their circuits to allow slight adjustments, so that the final circuit will behave as expected despite of all pre-production uncertainties.
 
@Theremingenieur .. nice thanks :) ..

should I have an eye on what tolerances the components i use in a ciruit have? maybe choose components with the same tolerance? .. of course there are not predictable things like copper traces, .. but, would that help or is that nonsense?
 
The classic way of dealing with these unpredictable effects in the analog world is :

1.) Build your circuit with the calculated components on a breadboard and do a thorough analysis of its behavior vs supply voltage, temperature, humidity, etc. variations until you find differences between the expected end the real circuit behavior. Identify the components or sub-circuits which are responsible for these deviations and either replace these by better suited ones before restarting the analysis or think about making one or more components variable to obtain the expected behavior. Consider that manual calibration is a huge cost factor in production and should be avoided if possible, through intelligent almost self-adjusting or self-compensating circuit design. Having done EE studies might be helpful here.
2.) Design a PCB with respect to thermal management and order a small quantity for prototyping. Build a few (10 or so) prototypes and analyze/observe these thoroughly as before, simulating all use cases and environmental conditions. Check if you can always obtain the expected results, if needed through one-time calibration of the variable components. If you need to recalibrate several times a day/week, it's poor design and you should go back to step 1.
3.) From these processes, you end up with an optimized circuit and PCB design and you might finally consider starting mass production.

There is no sense in choosing components with similar tolerances without deeper thinking and analysis. Lower tolerance means often higher cost, so the rule for tolerances is as low as possible, as high as needed. When it comes for example to a pull-up resistor for a switch, tolerances will not matter, take the cheapest one you can get. But when it comes to the generation of reference voltages, i.e. for precision analog signal processing in the uV domain in EEG or EMG applications, the more expensive 0.1% resistors might make things easier. Taking all that into account requires lots of trial and error at the beginning, but you'll acquire the needed experience over the years until you come to a point where you might do a big part of that work in mind, or where you look at some schematics and automatically identify critical components.
 
Status
Not open for further replies.
Back
Top