Electronics > Projects, Designs, and Technical Stuff

DC Bus Voltage sensing cirquit

<< < (4/4)

eliafavero:
I clearly did a poor job at explaining myself.
Find attached the mock up that I should have posted since the beginning.

I'll try to re-formulate what i want to achive. During the operation of the whole system it can occour that some of the many motors has to stop. (There are arround 200 motors connected to the DC Bus). Since the load won't be balanced anymore the tension V+ -> "0" and "0" -> V- (will abbreviate to V+ and V- for semplicity's sake) will have a different absolute value. If just a couple of motors stops, the shift in load is negligiable. If instead, an important amount of motors stops, we can come to a situation where we have for example V+ =300 and V-=240. The electrolytical capacitor on each Driving board ("drive" on the mock-up) has a maximal voltage of 350V. so if V+ is allowed to drift above that value it's easy to see what happens. BANG!
Up untill this point the originally proposed circuit could be used to do the trick and limit the voltage across each cap. Of course, there is a BUT...
During operation we will mainly use the motors as "motor" and not as generator, BUT the ultimate goal of the project is to be able to operate the drivers as bidirectional so that we can, occasionally, generate power back. Assuming we have an equal number of motor spinning on both sides and all working as generator, the voltage on the DC Bus will raise equally on V+ and V- e.g. V+=320 , V-=320. Now you see that if we simply put a limit on the maximum tension of each rail we would dump precious energy that could be otherwise be kept and used to re-accellerate the motors. Therefore what i actually need is some sort of circuit that keeps the DIFFERENCE between V+ and V- inside a certain limit.

hopefully this claryfies my point. I'll be glad to answer further questions.

Marco:
If efficiency matters the rail splitting is very counterproductive, you never did answer why all the motors couldn't be run from single rails.

If efficiency matters you should probably use DC-DC conversion for balancing.

If efficiency matters even when using resistive bleeding you should not try to preserve the mid point, you should start bleeding the higher voltage capacitor when the lower capacitor gets too low to operate the motors. Imbalance should not be an inherent problem assuming the motors have some kind of closed loop control, it only becomes a problem when the lower charged capacitor can no longer be charged from the rails.

eliafavero:


--- Quote from: Marco on August 22, 2019, 08:09:31 am ---If efficiency matters the rail splitting is very counterproductive ...

If efficiency matters you should probably use DC-DC conversion for balancing  ...

--- End quote ---

Efficency matters a lot for the project. The other thing that is even more important is the cost of the motors and drives. That is why we decided to split the dc bus so we could work with lower tension and have an easyer time designing the driving boards. Cost is also the reason why i am trying to avoid the DC-DC Solution.


--- Quote from: Marco on August 22, 2019, 08:09:31 am ---If efficiency matters even when using resistive bleeding you should not try to preserve the mid point, you should start bleeding the higher voltage capacitor when the lower capacitor gets too low to operate the motors. Imbalance should not be an inherent problem assuming the motors have some kind of closed loop control, it only becomes a problem when the lower charged capacitor can no longer be charged from the rails.

--- End quote ---

The goal isn't to preserve the exact mid point, infact we just want to limit the difference between V+ and V- by some amount (still to be determiermined). All the driving boards use a control loop to drive the motor.

Of course the goal isn't to operate a permanently unbalanced system constantly dumping energy. There is a central plc that comunicates with each drive and the plan is to stop the whole system if the unbalance persist after a set periode of time. this means that the balancing circuit will notify the plc when is dumping.

Ian.M:
I assume the only reason for the split rails is a motor voltage limitation + the desire to avoid a bulky transformer.
Even an autotransformer across the AC supply with a center tap to the DC 0V would keep the rails reasonably well balanced except during regeneration.

If you want to implement active balancing, drive a half H-bridge  at exactly 50% duty cycle to switch one end of a large, high current inductor between V+ and V-, with the other end tied to 0V.  If the voltages are balanced, the average DC current through the inductor will be zero.  If they are not balanced, it will transfer charge to correct the imbalance.  Its probably no more expensive than braking resistors as you are switching similar currents, but don't have to dissipate a lot of energy as heat.  There are a few more refinements needed - it would be advisable to add current sensing to take the half H-bridge hi-Z if the inductor current approaches saturation, letting the current flow through the MOSFET body diodes until its decayed enough to restart active balancing, and if more than half the motor load on can ever enter regeneration at the same time, you, probably still need a braking resistor across V+,V- to dump energy into, that cuts in at say 10% overvoltage and drops out at 5%, as previously discussed.

eliafavero:

--- Quote from: Ian.M on August 22, 2019, 08:54:09 am ---Even an autotransformer across the AC supply with a center tap to the DC 0V would keep the rails reasonably well balanced except during regeneration.

--- End quote ---

We are facing the unbalance issue exactly with the center tap...


--- Quote from: Ian.M on August 22, 2019, 08:54:09 am ---If you want to implement active balancing, drive a half H-bridge  at exactly 50% duty cycle to switch one end of a large, high current inductor between V+ and V-, with the other end tied to 0V.

--- End quote ---
That's quite a clever idea! This is the kind of circuit i am looking for. Did you just come up with this or have you employed the technique somewhere else?
I forgot to mention that the regeneration will only take place in case of a blackout. At that point all the motors (which all have the same load) turn into generator and the energy produced is directly used to drive some auxiliary motors. The regeneration and the correlated voltage increase is something i still have to investigate on, for now my focus is to get a stable and balanced DC Bus.
Thank you very much!

Navigation

[0] Message Index

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod