3 phase controlled rectifier with variable frequency 3 phase input

Thread Starter

ak52

Joined Oct 15, 2014
195
Hello everyone,

I have read the following thread and would like to take it one step further.
https://forum.allaboutcircuits.com/threads/3-phase-controlled-rectifier.65175/

My main question is how to you know when to switch on the SCRs if the three phase input in varying in frequency.

For example if the phase voltage has a frequency of 50 Hz, Each SCR must be fired for 60 degrees, so each SCR must be switched on for about 3.33ms.
How do i find out this value when the frequency is varying every cycle?

Thanks,AK
 

MaxHeadRoom

Joined Jul 18, 2013
19,041
The same way it is done in 1ph, the zero crossing point is detected and the firing point is either zero crossing switched or phase angle control is used.
Max.
 

ebp

Joined Feb 8, 2018
2,332
First, you only need three SCRs and three conventional diodes for a fully-controlled bridge. The linked thread talks about 6. The whole rectifier can be purchased as a module.

I did exactly this sort of things many years ago in a controller for charging batteries from a 3-phase variable frequency permanent magnet explosion-proof generator. I used CD4046 phase locked loops somewhat unconventionally to generate linear ramps synchronous with the appropriate half-cycle of each phase and compared the ramps and the output of the votlage error amp to generate the SCR firing signals. I would do it much differently now that there are so many fast, inexpensive 32 bit microcontrollers available. I'd measure the period using a timer and do the timing with either delay loops or additional timers. This requires division by 3, which was somewhat time consuming in the days of slow 8-bit controllers but is absolutely trivial now unless the frequency is very high. Doing the whole thing, including the error amplifier(s) in firmware is quite possible with a vastly reduced parts count relative to my original design.

If you only need to handle discrete frequencies (e.g. 50 Hz & 60 Hz) it is a simple task with even a very limited 8-bit microcontroller. I've used a small PIC processor to do the required timing in such an application for control of cathodic protection power supplies (external error amp & comparator - processor just did the timing for the phases & some supervisory tasks).
 

Thread Starter

ak52

Joined Oct 15, 2014
195
The same way it is done in 1ph, the zero crossing point is detected and the firing point is either zero crossing switched or phase angle control is used.
Max.
I seem to be missing some point here!,correct me if i am wrong,but unless i know the time period between one zero crossing to another(half the frequency of a cycle),then when i detect the zero crossing point,i calculate for how long should a pair of switches be ON.
So for example if the frequency is 50 Hz,and if i want to switch on a pair of mosfets for 60 degrees,then i can have a timer for 3.3ms and start the switching. If for the next cycle if the frequency changes to 48Hz,then the same 60 degrees, i will need to switch the mosfets for a longer time.

Or do i need some kind of phase locked loop to determine the phase before hand..you you pleae explain it to me further.

@ebp i am trying to do something similar,i and am trying to rectify a 3 phase input from a permanent magnet generator mounted on a modal wind mill.Could you please explain to me the algorithm in detail which you had employed.I will try and replicate it to me requirements,Thanks
 

MaxHeadRoom

Joined Jul 18, 2013
19,041
The way it is generally done with Triac switching is to detect the zero crossing point, many methods in use for this, and and then delay the turn on point that will give the required on-time until the automatic turn off at the next zero cross.
Max.
 
Top