hi,
This PDF is an introduction to jFET's.
E

Endless help, increadible I'm glad I found this forum, you guys are doing an awesome job!

 

Here's a substitute for the JFET that uses a TL431 programmable shunt reference to provide a more stable 4mA offset than using two diodes, whose forward voltage varies with temperature and supply voltage.
Q3's current is 2.5V/R4.



Edit: Here's a simulation of Eric's circuit with the TL431:

 

hi arrian,
It provides the 4mA offset current.
I will post an option.
E

Tweak the values to suit your App.

Hi, eric. In case of the schematic you posted, isn't 4mA current offset provided by Q3 dependent on +Vcc? That would be quite problematic.

 

hi arrain,
It is dependent upon the stability of the 12V supply, that is why the sim 12V is specified as 12.0v.
If the stability is a problem you could use say zener or the TL431 option shown in post #22.

E

EDIT:
This is with a 6v2 zener, add a 100n across the zener.

BTW: have you tested the performance of the 4mA source using +/-1V on the 12V supply.?
The change is not detectable.

 

hi arrain,
It is dependent upon the stability of the 12V supply, that is why the sim 12V is specified as 12.0v.
If the stability is a problem you could use say zener or the TL431 option shown in post #22.

E

EDIT:
This is with a 6v2 zener, add a 100n across the zener.

BTW: have you tested the performance of the 4mA source using +/-1V on the 12V supply.?
The change is not detectable.

You're right, I thought changing Vcc would mess up more. Anyway, zener diode works pretty well. Final design in the attachment.

 

Hi, I am finishing this project, about to design pcb. I have some concerns about power supply through virtual ground. I need to Supply Raspberry Pi with 5V. I'm using LM7805 (in series with 12V one to reduce overheating) to do so. RPI will draw approximately 1A, at max output at both channels the current draw will be around 50mA at 17V = 0.85W. Is putting 10W zener diode in virtual ground circuit enough for providing sufficient current supply? Schematic in attachment.

 

hi arrian,
Is the section of the power supply you plan to install a 6.8V 10Watt Zener.? If so, I don't see any current limit resistor.??

E

 

hi arrian,
Is the section of the power supply you plan to install a 6.8V 10Watt Zener.? If so, I don't see any current limit resistor.??

E

View attachment 177882

What about R2?

 

hi,
What will be the current loading on the +v and -V
Look at this Sim, using 1K as dummy loads.

What do you expect the current to be thru the 10Watt zener.?

E

 

I use the AD5412 chip.
https://www.analog.com/media/en/technical-documentation/data-sheets/AD5412_5422.pdf

 

hi,
What will be the current loading on the +v and -V
Look at this Sim, using 1K as dummy loads.

What do you expect the current to be thru the 10Watt zener.?

E

I expect around 1A current draw from the LM7805, which means around 0.3A draw from +Vcc and zener diode at 17V.

 

I expect around 1A current draw from the LM7805, which means around 0.3A draw from +Vcc and zener diode at 17V.

Unless you are using switch mode regulators, the current from the 17V supply will be 1A as well. Actually, even higher if you are extracting current from the 12V point. Linear regulators input current is the same as output current, ignoring the small control current. So heatsinking will be needed. A switch mode supply will save the heat and the current from the 17V supply will drop significantly.

 

hi arrian,
I agree with 'dendad' point, use a SMPS, your present circuit is very inefficient.
E

 

hi arrian,
I agree with 'dendad' point, use a SMPS, your present circuit is very inefficient.
E

I see, switching regulator will be a better choice. What chips do you recommend? Will eg. AOZ1282CI or AP5100WG-7 be suitable for this application? Or maybe fixed-5V one like this: IFX91041EJV50?

 

hi arrian,
Have you considered a ready built SMPS module, from eBay or Amazon.
The IFX d/s is attached.
E

 

These seem to work well...

This is an Australian Ebay advert.
Some of these have solder links to set the volts, while others use a trim pot.

 

hi arrian,
Have you considered a ready built SMPS module, from eBay or Amazon.
The IFX d/s is attached.
E

Yes, I mentioned it in my question. I guess It will fit

 

hi arrian,
Have you considered a ready built SMPS module, from eBay or Amazon.
The IFX d/s is attached.
E

It seems like IFX91041EJV50 is not quite popular. The datasheet briefly describes the typical application for fixed output, but some things are not clear for me. What is max EN enable voltage, can it be connected directly to +Vcc (17V)? What value should D7 have? Is it supposed to short over-voltage spikes, being eg. 10V schottkey diode? I know It is probably somehow written in the d/s but I want to clarify my concerns, since there are no egzamples on the web.

 

I'm not familiar with the IFX91041EJV50.
The boards I've used are as above, and a similar one with the MP2307 chip.


The resistive divider, R1 and R2 set the output voltage. That will be the same for the chip you are looking at. If the FB pin is just connected to the OUTPUT, the volts will be the chip default voltage. Note, some chips come with internal settings for fixed voltages, while they are also available in an "ADJ" version.

 

Hi. Since my last reply, I have finished the design and ordered the pcb. I'm now during debugging stage, already fixed some mistakes but there is something more wrong I can't figure out. I attach a pdf with complete schematic, and screenshots with the actual wiring (changed for the purpose of debugging) with expected voltages. I don't know why U4B gives different voltages at the inputs and why there is 5V at the input of U4B (5th pin). Everything seems random, is it a broken chip?