How much current needs to be drawn to pull DALI bus from 16V to 0V?

[ocset]

Hello,
To be able to transmit on a DALI bus, the TX must be able to pull it down to 0V. How much current must be drawn from the DALI bus to pull it down to 0V?

[rstofer]

Here are the specs:
https://web.archive.org/web/20130627012349/http://www.dali-ag.org:80/c/manual_gb.pdf

The master control unit needs to provide 250 mA (actually 2 mA/device) and the voltage ranges are easier to find in the Wiki:
https://en.wikipedia.org/wiki/Digital_Addressable_Lighting_Interface

[ocset]

Thanks, so if there's only one lamp on the DALI bus, then it has to be able to draw the full 250mA out of the DALI bus in order to make a low pulse?

[rstofer]

The way I read the spec is that each unit is 2 mA and the MAX on the chain is 250 mA.

[ocset]

Thanks,
Does DALI require the micro in the lamp (the receiver of DALI signals) to have an oscillator frequency of 16MHz?
If not, how low can the osc frequency be for receiving DALI commands? (ie the osc frequency of the micro in the lamp)
Also, does DALI need the micro in the lamp to R/W to EEPROM?...our software guy says it does.
Also, does DALI need the micro to have a random number generator function?...our software guy says it does.

[Kjelt]

- yes the device has to be able to sink 250mA on the bus and it is only allowed 2mA to draw from the bus in order to operate

- Your micro has to reliable decode 1200 bits per second, usually 4 or 8 times oversampling is used or irq's with timing functionality.

- When the device is commisioned it has an ID, that has to be stored somewhere, could be in eeprom or fram, I would not use rom for that.
  Other parameters you might want to store like groupid, last dimlevel etc could also be usefull to store in eeprom.

- a real true HW rng is not necessary in software you can build a nice pseudo-rng which would suffice but actually DALI needs more an unique number generator than a random number generator

[ocset]

Thanks,
Though the circuit on page 3 of the AN1465 app note by microchip wouldn’t be capable pf drawing 250mA   from the DALI Bus if the source driving the TX opto was only  driving with 2.5V when high. Do you agree?

This is because the MMBT2222 transistor only has a hfe of about 30 when VCE = <1V.

The TCLT1000 opto has a CTR of 0.5*0.6 at the minimum.

TCLT1000 opto:
https://www.vishay.com/docs/83515/tclt1000.pdf

AN1465 app note:
http://ww1.microchip.com/downloads/en/AppNotes/01465A.pdf

MMBT2222:
https://www.diodes.com/assets/Datasheets/ds30041.pdf

[Kjelt]

I do not know that circuit, I do know that many circuits from uC manufacturers as Atmel/MC and ST in their Apnotes will not 100% pass the standardised Dali testers, so figure.
Some companies have very strict additional demands for the circuit such as that it should survive a wrongly inserted mains connection, so it should survive 240AC on the Dali input if the installer made a mistake.
If you want to design something for the professional market my suggestion is to buy some decent brand like Osram or Philips Dali drivers and look what is in it.

Looking briefly at the circuit, how did you come up with the 2,5V?
It is a 5V logic system, the GPIO port will be driven with 5V high output.

[ocset]

Thanks, our microcontroller in our lamp has a Vdd of 2.5V.
As such, we will be pulsing the opto with the 2V5.

[ocset]

Thanks, as you can see, it cant be so readily be done with a FET....the component count woudl be higher.
The attached ltspice sim shows how it doesnt work so well with a single fet.