Filter Caps 3

[CBravo]

How can I calculate the current draw of a filter capacitor bank.
We have a customer who wants us to electroplate at less than 5% ripple. We are at about 30% ripple now. We plate at 6.5 DC volts and 11,000 amps with 2.5 AC volts ripple. So how much capacitance would be needed to smooth it out to 5% and would the bus to the cap bank need to handle 4000 amps?

I'm trying to work out a ballpark cost of this mod.

Thanks.

 

[electricuwe]

Regarding reducing the ripple of a plating rectifier the design of the rectifier control is important. Basicly there are three different designs:

- with variable voltage transformer
- with primary thyristor control
- Switchmode Power supply

Judging from your current ripple condictions statet above I guess you are using a unit with primary thyristor control.

To limit ripple in such a unit a DC smoothing choke is a more approbiate way to reduce ripple compared to a capacitor. For choosing a suitable choke you should consider if you will always use the full current or if also wnt to plate with low ripple at reduced current. This will have significant impact on the required inductance.

In the past I designed such units as as oil cooled units with the choke included in the tank. For an retrofit application this is unfortunately not an option.

A separate choke would be my choice, if the rectifier is in good condition. I would prefer to buy a completely new unit, if the rectifiers expected lifetime is limited due to wear or unavailibility of spare parts / qualified service.

 

[Skogsgurra]

It sounds like we have been working on the same unit, Uwe.

I have also had a few thyristor controlled units with analogue gate control. The excessive ripple was, in this case, due to bad firing pulse symmetry. It took me a couple of days (no diagrams available) to adjust the firing unit so pulses were back on 60 degrees equal distance across the whole control range.

Find out if bad pulse timing is the problem or if you really need more smoothing before you do anything. If bad timing is the problem, it is usually a good idea to get a digital firing unit. The rectifier as such is usually OK.

Considering the problem to find anyone that understands analogue control, it is also a good idea to get more modern gear.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[electricuwe]

Gunnar is right in stating that bad symmetry in pulses causes additional ripple, but a plating rectifier with primary thyristor control will also have significant ripple with perfect pulses when operated with less than nominal voltage.

CBravo please provide nameplate data of the unit under consideration.

Plating rectifiers with primary thyristor control are significantly less expensive compared to variable voltage transfomer designs if no choke is beeing used. But at voltages lower than nominal, thyristor control without choke leads to much higher ripple compared to the classic variable voltage transformer design.

 

[CBravo]

The unit in question is a 12 volt 25,000 amp Dynapower rectifier. The rectifier is at our other plant, so I don't have all the nameplate data yet. It is thyristor controlled. The load varies depending on the part we are plating, but generally the rectifier is loaded up to around 13,000 amps at 9.3 VDC. I measure 2.6 VAC on the bus. The next time I'm there I will have a look at the waveform on the scope to see how the firing symmetry looks. I am almost certain it is analog control. I will check that as well.

Thanks Guys.

 

[electricuwe]

I found one paper desribing the topic in detail, unfortunately in german, but the figure on the last page shows what I explained in earlier posts.

Usually all three-phase plating rectififers use 6p rectifier circuits.

http://www.munk.de/downloads/restwelligkeit_hc.pdf

 

[CBravo]

Good stuff. I wish my German was better.

I was able to look at the waveform on a similar plater with a scope. I was pleasantly surprised to find the ripple voltage less than 5%. This is due to a huge external choke. This works very well.

The other plater is not so equipped and has twin 12000 amp rectifiers that are usually run in parallel to achieve sufficient current. This system has a typical sawtooth waveform with 30% ripple. It would be very expensive to add a choke to this system. A filter capacitor retrofit might be practical if I could calculate the necessary capacitance.

 

[CBravo]

The translation into English.

 

[ScottyUK]

CBravo,

The following might help. There are some example calcs halfway through.

http://www.evoxrifa.com/electrolytic_cat/electrolytic_appguide.pdf

I don't think a capacitor solution will be economical in comparison to a choke - the physical size and cost of a bank large enough to handle the ripple current you are considering will be very significant.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[CBravo]

Scotty,
Thanks very much. That was exactly what I was looking for.