Everything goes from analog electronics, 3d printing, microcontrollers, embedded systems, imaging, AI, data visualization, electric cars, rockets, weapons of mass destruction etc...
Basic attitude of this blog is "FUCK YOU", "FUCK EVERYTHING". Nothing is holy or totem animal. :D :D This is like how if Donald Trump would be an engineer
This post idea started from my previous ranting post about reliable electronics. And how reliable looking device can be in reality just piece of shit...actually highly overprized piece of shit
http://hasseb.fi/shop2/index.php?route=product/product&product_id=56
This is no offence Don't get me wrong. My plan is to rant about bad designs and show how to real man do the job. I will provide electrical and mechanical designs for cheap,compact general purpose current to voltage amplifier module. and publish design as open source hardware.
All of them
- Schematics
- Layout (for cheap small two layer pcb)
- Bom, bill of materials
- Some good choices for mechanics (enclosure, connectors etc..)
- Directions where purchase parts
Schematics and layout are provided in kicad format and in pdf and gerber formats. Not the fracking propietary eagle.
Very low current to voltage amplifier is so essential tool for so many fields of sciences and engineering. So I am amazed why there are no cheap solution this problem.Only bad designs
Bad design is usually caused by
- Designer who does just what is asked and nothing more. And do not look big picture or do over engineering
- Boss who is not technical enough and do not have a vision
- Stupid customer is clueless. Dont know anything or have a pervert vision about what he/she needs.
- Designer who do not have enough confidence, vision and profesionality to say loudly and clear "go fuck yourself" to boss and customer. And after that make a great product or service.
My goal is to create situation where it will be possible to purchase cheaply low current amplifier as complete or as kit. (because I also need amplifier) And it would be very easy to connect that anywhere. This will enable researches and hobbyist to do all kind of experiments that they were not able to do earlier.. Paying for over 2000Eur for amplifier vs around 100Eur for "complete" unit will generate so much good things. (also others than myself) Even small scale instrument manufacturers like hasseb will generate revenue.
Interfaces and interfacing
How this amplifier system is interfacing. And I am not talking about connectors. Required connectors are connected to board.. directly or by wire. Thats mostly mechanical engineering.
Another side of the coin is the electrical interfacing.
Now the power comes from wall. AC voltage, transformed down, diode bridged and regulated to +12v -12v rails. And
Problems with electrical interfacing
- AC power issues
- Safety? If done right
- AC brings 50Hz or 60Hz hum to system. 50Hz lies on measurement frequency band
- For extra quiet measurements battery would be better?
- Integrating amplifier to part of larger system is hard. It sucks to handle high voltage AC
- Linear power supply
- Two models for transformer 115V and 230v,
- IT is so 1900's to have to be carefull about PSU switch position etc..
- Logistic nightmare if selling world wide
- Accidents when US customer brings amplifier to europe
- Efficiency.
- Linear regulator is just "smart resistor" that adjust its resistance so output would be the at "setpoint"
- voltagedrop*current = that fucking heat power. ALWAYS with linear regulator. (Regulator might say 1.5A, but watch the power)
- That fucking heat must be transfered away from regulator btw
- Obsolete opamp?
- High operating voltages?
- +12v and -12v rails
- Any extra power-rail increases total costs (parts+complexity+manufacturin) of full system
- Powering from one battery?
- Better performance opamp available?
- Where output is connected?
- In 2010's the output result is eventually converted to digital result.
- The most of ADCs works from single side side supplys
- Negative voltages are big no-no
- Can be overcome by level translator resistors/buffers and maybe protection diode+resistors
- Most ADC's max absolute (measurement range is different story) is related supply voltage
- +5V, like arduino pin
- +3.3V like cheap ADC running 3.3v
- +1.8V like beagle bone ADC
- Today measurement system must be designed so results are converted and transfered to computer system at as early stage as possible.
- Wiring (electrical wires are related to electrical interfacing)
- Why two separate boards?
- Why there are so much wasted board space?
- Are pcb connectors really so expensive that 2500USD instrument can not have them?
- Solder wires to pcb and if wire is shaking... wire will break eventually from where solder ends. Or thin conductor trace peels off.
- Why gain selector is connected in between -IN and feedback resistors?
- Input is high impedance,.. and high impedance net goes to wire, wire to switch, switch to wire. Too many places where noise can couple capacitively
- Output is low impedance and it just goes to resistors
 |
| WHY THE HELL THIS IS DONE LIKE THIS??? How about if opamp drives gain switch and resistors are nearby +IN.... It is so basic eletronics 101 |
Better interfacing
One size does not fit all, but making amplifier easily modifiable is must have feature. Just two solder jumper spots and magic happens.
 |
| Red dots are soldered jumper bridges, ust quick idea how powering and connectors should go |
- raw power input
- current input
- current in wire
- possible to connect: virtual ground or "most negative"
- power in/out and signal
- get regulated +5v and +2.5v (virtual ground) for external circuit
- disable regulator, but feed power from "power in" cable
- Normal: Board is powered from DC-power source.. like 9v or 12v battery
- jumpers are closed
- +5v and +2.5v volts are available for external usage
- Can measure positive and negative current
- Powered completely from opamp power lines
- both jumpers are open
- potential in negative and positive wire must not be over operational voltage tolerance (max +5v)
- Positive only: Powered by DC power
- +2.5v jumper open.
- opamp gnd short circuited to "the most negative"
- output goes 0-5v, only positive current
- Positive only: Powered by "opamp +V"
- Both jumpers open
- opamp gnd short circuited to "the most negative"
- output goes 0-5v, only positive current
- Powered from external low noise +5v power rail
- +5v jumper open
- opamp gnd is left unconnected
- The "most negative" is connected to external ground
Grounding and shielding
"When designing a satellite .. the most important thing in system integration is proper grounding plans, even grounding is very silly named thing in satellite"
-Origin is unknown
When using high gain, it will be good idea to have metal enclosure. The main disturbance coupling mechanism is capacitive. Capacitive coupling can couple only alternating voltages (unless capacitor electric field sucks charged particles from somewhere). It means that potential difference in between metal enclosure and opamp "gnd" (virtual gnd or bipolar power supply gnd) must be constant. It is ok to connect enclosure to virtual "gnd" or the most negative rail
So there MUST be connection in between pcb power plane and metal enclosure.
The planned point of grounding would be from maybe at input (input wire is coaxial) or externally
PCB is not going to have grounded mounting holes. This will allow system integrator to have control where and how return currents are returning. :)I will provide more detailed and user friendly instructions when circuit is done.
Next part
On next part I will investigate possible component selections and design choices.Stay tuned. Next part maybe at next week
Ei kommentteja:
Lähetä kommentti