Connecting the DC-bus of 2 VFDs.

JesperMP

Lifetime Supporting Member + Moderator
Join Date
Feb 2003
Location
ᚴᚬᛒᛅᚾᚼᚬᚠᚾ
Posts
16,239
Hi !

I have an application where a vibratory conveyor is driven by 2 agitators.
In order to change the transport direction slightly (for sorting the material on the conveyor), the agitators are driven by VFDs that drive the motors at a slightly different frequency (f.ex. 49.5 Hz and 50.5 Hz). That cause the agitators to run at the same speed (due to the mechanical coupling), but a difference in the angle, which in turn causes a slight angle of the conveyors transporting direction.
Since one agitator is essentially driving and the other braking, the braking VFD has to send a lot of energy to a brake resistor.
This all works OK, but the brake resistor has to be much larger than normal, since it has to bleed off energy constantly rather than only when stopping.
This in turn also means I do not have much room for adjusting the frequency, since the resistor may overheat (despite being dimensioned much greater).

So I am pondering if it could pay off to connect the DC-bus of the 2 VFDs.
The idea being that the braking VFD sends the energy to the driving VFD, and thus do not have to bleed it off in the resistor.
I have heard of this principle, but I have no experience, and I would like to hear from anyone with experience or links to instructions regarding this.
Anything to be aware of ?
The VFDs are Siemens G120. Do some parameters have to be adapted when connecting the DC-bus ?
 
Last edited:
I am probably not understanding the problem correctly, but why would a different supply frequency cause the motors powered by VSDs to run any differently? There isn't any direct connection between the speed of a VSD supplied motor and the frequency of the supply.
 
The VFDs output frequency to the motors is different.

I guess the way I phrased it, it is not clear that I meant the supply to the motors, not the VFDs.
I have edited the 1st post to make it clear.
 
I have experienced the downsides of connecting the DC bus of multiple VFDs, the end result is quite catastrophic, and the smell (aka "the stench of failure") stays with you for a long time. The big question that the drive manufacturers seem to forget, if there is a fault on the DC bus of one drive, how will that energy be dissipated before it reaches the next drive.

Case 1: Machine with 10 VFDs connected to a common bus, no protection between drives. 24v DC power supply was undersized, under certain conditions 24V supply would dip enough that drives would go into fault and all try to discharge their capacitors through the same DC bus. First time it occurred, 8 dead drives. We stopped counting after 30. Luckily machine was under warranty, but when the manufacturer is delivering VFDs by the pallet via airfeight, it's time to rethink the design.

Case 2: Machine with 2 VFDS sharing common DC bus, this time seperated by high speed semiconductor fuses. Unfortunately, fuses were not fast enough, 2 dead drives instead of one, and then establishing the root cause of the fault becomes signifcantly harder.

On the flipside, we have 4 other machines from the same manufacturer with same series of VFDs mentioned in case 1, each machine has 3 drives sharing a DC bus with no segregation between drives, and we haven't has a problem.
 
I have done a common DC bus link similar to the 'Chapter 6 - Piggy-back configuration' example in the Rockwell document that was previously posted. This setup was with 2 ABB ACS550 drives. One drive with AC supply, then the 2nd drive was connected via the DC bus.


On this particular machine there is a large (70hp) driven motor and a smaller (20hp) braking motor. It was known that that the smaller motor would always be in regeneration, but regardless we sized the larger drive for 90hp to ensure the 'front end' of that drive could handle any possible load it would encounter.
We used a high-speed DC fuses in the bus link.
 
gcc, I will probably install semiconductor fuses in the DC circuit.

JRW,
6SL3224-0BE27-5AA0 PM240
6SL3244-0BB12-1FA0 CU240E-2 PN
6SE6400-4BD16-5CA0 Brake resistor
 
The drive knows when it is braking, isn't there anyway to use that to trim the speed of the drive?


edit:

Hadn't though that through, you would have to trim to speed of the second drive.
 
The two agitators must be driven with different frequencies to get the desired effect.
It is impossible to not have one VFD braking when the other is driving.
 
Thanks guys.
I am studying these right now.

1st question. In the http://literature.rockwellautomation.com/idc/groups/literature/documents/at/drives-at002_-en-p.pdf, on page 54 (and other pages) what does the "Back-to-back diodes that are required for proportional current sharing" actually do ?

Check out pg. 133. The manufacture of the modules has some good info on what they do. Here is a manual for one of them - http://www.bonitron.com/PDFs/Manuals/D_M3345D_CMAN_VALL_02f.pdf
 
Check out pg. 133. The manufacture of the modules has some good info on what they do. Here is a manual for one of them - http://www.bonitron.com/PDFs/Manuals/D_M3345D_CMAN_VALL_02f.pdf
The schematics in the AB PDF just says the back-to-back diodes are needed. I cannot see what the illustrated parallel diodes in opposing directions actually do.
The schematics in the bonitron pdf page 15 is different to AB, but it makes sense to me.
I am guessing that bonitrons description is the correct one.

edit: No wait, the bonitron illustration does not apply to my case.
I have several VFDs as in the AB PDF page 54, not with socalled "regen modules" as in the bonitron PDF page 15.
I am still mystified as to what this diode module does, and how I should apply it in my application, and if I even need it.
 
Last edited:

Similar Topics

I’m attempting to configure KEPWare to poll a Modbus device via Modbus TCP/IP. However, everything I’ve tried results in an error being reported...
Replies
2
Views
373
I am trying to get a Mitsubishi GT2715 HMI to read and write from a Eurotherm 3508 controller using 4 Wire RS485 Mobus RTU communications. The...
Replies
3
Views
1,285
I'm trying to resurrect an older machine from around 2005 that uses a 140CPU43412A. The goal is to retrieve the existing program so that it can be...
Replies
4
Views
3,344
Hi All, I have a closed loop stepper motor (Nanotec make), https://en.nanotec.com/products/2512-pd4-e601l42-e-65-4. This uses MODBUS TCP for...
Replies
4
Views
3,698
Hello, I'm Alex from Spain (sorry for my bad english). Firstly, I'm a newbie at Siemens PLCs and HMI, I have been reading this forum for hours...
Replies
2
Views
1,570
Back
Top Bottom