+ Post New Thread
Page 2 of 3 FirstFirst 1 2 3 LastLast
Results 21 to 40 of 56
  1. #21
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Vicor Bus Converter module

    Its a power supply for a 12kW drone.
    If a primary switch fails, (ie the high side one in the buck(s)) then the drone emergency lands on its reserve battery supply.

    - - - Updated - - -

    Aaah, I see what you mean...the high voltage would then get through and damage the motor/drive......that's why you recommend full bridge?

    - - - Updated - - -

    ...though we could just use the Bucks anyway and somehow crowbar it if that happened...and save the motors.
    Last edited by treez; 11th February 2020 at 09:24.



  2. #22
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Vicor Bus Converter module

    Hi
    Unfortunately there’s no chance of splitting up the power supplies, ie one per drone motor…so its one single big power supply. That’s what’s been decreed…

    To re-iterate, The power supply has 820VDCin
    Load = 50V at 240Amps (12kW)
    There is a 320 Metre long cable between 820V source and load. (9.5 ohms go and return)

    We believe that there is a possible way that this can be done with Vicor modules….

    It involves one Vicor BCM4414xH0E5035yzz module feeding four paralleled Vicor DCM3623T50M53C2T00 modules……the output of the four than goes via a current monitor to the 240Amps bus…and 12 other lots of the above would also feed into the same bus.
    (datasheets for these modules are in post #14 above.)

    The schematic is as attached. (just shows one “group”, not all 12.)
    The current monitor output would be fed to an error amplifier, which would ensure that no group of four DCM.. modules fed more than 24 Amps of current into the common bus…….in this way we would get “droop” sharing, and no significant current hogging …the error amplifier would be fed, as in the schematic, to the TRIM pins of the DCM.. modules

    We think this overcomes any significant sharing problems.
    Can you see any Gremlins in this?

    -----------------------------------------------------
    I must confess, I would like to address this with eight paralleled, synchronous 1.5kW two transistor forward converters (Sic FETs throughout)….all being fed the output of a single external error amplifier, which would regulate the single Vout….but we are bound to Vicor, as its perceived to be quicker ,and more commonplace tech with a globally proven track record.



  3. #23
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Vicor Bus Converter module

    ...on the theme.......we could have eight identical paralleled 1.5Kw two transistor forwards……..and put a current limiter on the output of each one…….so that those that hogged hit their current limit, and then the others start coming in……..”droop” sharing.
    So we want 50V at 240A in total…….so for each two transistor forward we would add a 30A current limiter to its output and then the whole eight of them would share with no other circuitry required…and their feedback loops would not fight each other....would you agree this is the simplest and cheapest way to do it?



  4. #24
    Super Moderator
    Points: 54,106, Level: 56

    Join Date
    Apr 2011
    Location
    Minneapolis, Minnesota, USA
    Posts
    13,188
    Helped
    2628 / 2628
    Points
    54,106
    Level
    56

    Re: Vicor Bus Converter module

    It's a good idea to anticipate that one power supply will conk out for whatever reason. Immediately a group of propellers loses power. What the craft does next depends on the arrangement of motors:

    (a) If the dead motors are at one corner of the craft, that corner drops (perhaps drastically),
    or
    (b) if the dead motors are in varied locations around the craft, then the craft may remain level.

    There is an advantage to having independent control to deal with the unexpected. You may need to regain level flight, or stop it from oscillating, or stop it from rotating, or return it to base quickly. Etc.


    1 members found this post helpful.

  5. #25
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Why dooes offtheshelf DCDC module limit the numbr of parallel units?

    Hello
    Page 1 of the DCM3623T50M53C2T00 datasheet states that only up to 8 modules can be placed in parallel.
    Why is this? Why not say, 40 in parallel?

    DCM3623T50M53C2T00 datasheet...
    http://www.vicorpower.com/documents/...53C2yzz_ds.pdf
    Last edited by FvM; 23rd February 2020 at 18:42. Reason: Vicor threads merged, Crossposting of previous discussed question



  6. #26
    Advanced Member level 3
    Points: 7,012, Level: 19

    Join Date
    Feb 2014
    Posts
    998
    Helped
    349 / 349
    Points
    7,012
    Level
    19

    Re: Why dooes offtheshelf DCDC module limit the numbr of parallel units?

    Where do you see '8' referenced?

    Regardless I assume its not a hard limit but rather it's based on practical difficulties of implementation - like ensuring that routing is well matched between all 8.


    1 members found this post helpful.

  7. #27
    Advanced Member level 2
    Points: 5,454, Level: 17
    Achievements:
    7 years registered
    ZASto's Avatar
    Join Date
    Jan 2008
    Location
    Messy Corner, 44°47'31.56"N, 20°28'9.16"E
    Posts
    661
    Helped
    137 / 137
    Points
    5,454
    Level
    17

    Re: Why dooes offtheshelf DCDC module limit the numbr of parallel units?

    As @treez wrote on tve very first page of attached .pdf

    Click image for larger version. 

Name:	Screenshot from 2020-02-23 17-22-56.png 
Views:	5 
Size:	83.2 KB 
ID:	157916
    Make no mistake between my personality and my attitude.
    My personality is who I am.
    My attitude depends on who you are.


    1 members found this post helpful.

    •   AltAdvertisement

        
       

  8. #28
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Why dooes offtheshelf DCDC module limit the numbr of parallel units?

    Thanks, the routing situation is what i first thought, but the DCM3623T50M53C2T00 module uses "droop" sharing....(its vout decreases by some 5%, from 10% load to full load)......so routing differences shoudlnt matter with this module
    (in fact, it will matter even less with us because we have implemented an external error amplifier to make them all current limit at 6A, so no one DCM module can hog the current.).
    Its almost as if Vicor dont want to sell many modules?



  9. #29
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Vicor Bus Converter module

    Why does the vicor DCM module have unlimited output capacitance allowance (well its limited to 2mF but that’s only due to startup time, not due to instability)…but the Vicor PRM module must have a limit of just 47uF for its C(out)?
    None of the vicor literature states why. What will happen if we violate and use say 3mF on a PRM output?

    Vicor DCM
    http://www.vicorpower.com/documents/...53C2yzz_ds.pdf

    Vicor PRM
    http://www.vicorpower.com/documents/...T400A00_ds.pdf



  10. #30
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Large numbers of paralleled DCDC modules?...more than datasheet recomends

    The Vicor DCM* module (DCDC converter) datasheet (page one) states that up to 8 DCM* modules can be put in parallel.
    But What if we put 40 of them in parallel?
    (48vout, 240Aout)

    What will be our downfall?
    1…Beat frequencies between modules causing enormous input and output overvoltages?
    2….Enormous ground current loops resulting in noise susceptibility such that the DCM* modules literally just stop working?
    3….Just a general high ambient noise level, caused by 40 SMPS modules operating in reasonably close proximity…..as such, I take it that common mode chokes to the input of each DCM* module would be mandatory?
    4….What else?

    What is the probability of it being fine?
    The 48Vout is just a rough input voltage to a couple of BLDC inverters.

    DCM* module datasheet:
    http://www.vicorpower.com/documents/...53C2yzz_ds.pdf


    ....................*........................*.... ........................*........................* ............................*..................... ...*............................*................. .......*........
    Also, what if we (purely for sake of argument) buy 40 Meanwell LRS-350, 350W Offline SMPS’s, and sit them all next to each other…and connect their outputs to their own 300W dummy loads? Also connect them to the same mains cable (assume it can handle the 12Kw). Will there be noise problems?...
    LRS-350
    https://www.meanwell.com/webapp/prod...x?prod=LRS-350

    - - - Updated - - -

    We assume that having a common mode choke in front of each DCM* module will greatly improve the chances of success of having 40 DCM* modules in parallel?
    Last edited by treez; 1st March 2020 at 18:42.



  11. #31
    Advanced Member level 5
    Points: 13,285, Level: 27

    Join Date
    Aug 2015
    Location
    Melbourne
    Posts
    2,180
    Helped
    795 / 795
    Points
    13,285
    Level
    27

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    a small load sharing R on each module should allow sharing up to N modules, just be sure the input wiring is symmetric the source can supply the power and the output leads are reasonably symmetric too ...


    1 members found this post helpful.

  12. #32
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    Thanks, i think we will struggle to achieve all the symetry...therefore, what we have done, is put an individual current clamp error amplifier on each individual DCM* module....so that no module can supply more than 6.5A........in this way....there will be far less hogging of current...at least, that is our intention.

    The attached shows the current clamping for each DCM* module...only one DCM* shown here.



  13. #33
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    Please may I first state that Vicor DCDC modules are superb beyond belief…they are the best…we all know that …there are probably billions of them out there doing their wonderful job……

    The thing is, the Vicor DCM datasheet states that there is a limit on the numbers that can be paralleled. …now, in the following thread, there was a general concensus that a virtually unlimited number of power supplies of they type described there could be paralleled, in the way described there……so why is it that unlimited numbers of vicor modules cannot be paralleled? (why are we limited to 8 vicor dcm modules in parallel?, as the datasheet says)

    https://www.edaboard.com/showthread....h-power-SMPS-s

    …..is there a situation with vicor modules, that they are more susceptible to noise than standard power supply designs?……I say this because the vicor modules are extremely small…their power/square_inch is unbelievable……but just maybe this comes with a downside…..that downside being that compressing a load of power and signal circuitry so closely together means that the module is somewhat overly sensitive to noise issues?.....
    To make matters worse…the vicor dcm modules, have a crossover frequency of 20khz…very fast…and we all know that high crossover means more susceptibility to noise.
    I am just thinking that if we do a standard Two transistor forward design, (like in the thread linked above), then we may be able to do loads of them in parallel and get to the required high power level (12kW)….something that it appears we will not be able to do with vicor modules....as per vicor datasheets.
    I suspect that if we did a standard two transistor forward design, with output current limit, then we would be able to put loads of them in parallel and achieve our 12Kw……the above thread appears to agree with this. Do you have considerations.



  14. #34
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    Hi,

    Despite the Vicor DCM module datasheet stating that no more than 8 DCM modules can be put in parallel….the below vicor app note shows 18 vicor DCM modules being put in parallel.

    This suggests that maybe one can purchase “modified DCM” modules that can be paralleled in large numbers (>8) …..do you know of this?

    Perhaps the ones that can be put in parallel arrays of more than eight have had their feedback loop compensation components adjusted so that the loop is slower and thus giving a not so noise sensitive DCM?...and/or the damping of the fet switching has been increased in order to decrease the noisiness of the DCM? Perhaps even internal shielding was used in a "modified DCM" structure , in order to get the parallelability up to 18?


    DCM3623T50M53C2T00 datasheet...
    http://www.vicorpower.com/documents/...53C2yzz_ds.pdf

    Vicor app note showing 18 DCM DCDC modules in parallel for a 9Kw load
    http://www.vicorpower.com/documents/...se_Vehicle.pdf



  15. #35
    Advanced Member level 5
    Points: 13,285, Level: 27

    Join Date
    Aug 2015
    Location
    Melbourne
    Posts
    2,180
    Helped
    795 / 795
    Points
    13,285
    Level
    27

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    As stated previously above - there is no obvious point against having many modules in parallel as long as each module can see a low Z supply on its input - and there is a modicum of small sharing R on the output, such as provided by an LC filter ...


    1 members found this post helpful.

  16. #36
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    Thanks, while I am absolutely certain that you personally are capable of designing power supplies that can be put in parallel in extremely large numbers (>>8) ……in many ways, including the way that you state above……..i suspect doing it with vicor DCM modules presents problems…..otherwise I just cannot believe they would have this in their datasheet that the Vicor DCM modules can only be paralleled to a maximum of eight….this has been in vicor datasheets for years. If they could go more than eight, they’d say so, simply because it would mean more sales.

    I believe part of the problem of >8 is that the crossover frequency is so high (20kHz), and also, there is probably little damping of the gate turn ON and OFF, and though its ZVS…..there’s probably a hard turn off, and maybe not even a completely zero voltage turn ON……there’s possibly also little internal shielding within the modules....which are extremely compact, with power and signal circuitry crammed tightly together..........not a recipe for success when you are shovelling loads of them tightly together in parallel...even with low enough impedance filtration at the input, etc.

    Oh, also, I appreciate your comment about R at the outputs……I’m not being funny or anything, the DCM modules actually have a negative load line and their Vout drops by 5% from no load to full load…….so this , as their datasheet says….(or one of the app notes?) means no series R is needed at the output. ( as you know, you’ve actually kindly already recommended such a technique in one of your posts, though not with reference to the DCM).

    Reference the low Z supply at the input…I’ll say…vicor are very vocal about this…….the input filters output impedance must be 10 x less than the DCM input impedance at all frequencies up to crossover (20khz)………also, the f(sw) of DCM is 750kHz, so they are not going to see an electrolytic capacitor….so its not like we can just shove some big electro’s on at the input and forget about it….….so we have to use ceramics…………..then we have the problem that in order to stop disturbances between paralleled DCMs, we have to have an input filter inductor (as per datasheet)……which then rings with the ceramic capacitor input bank……..and means that achieveing the “10 x less” impedance of the Zout of the input filter is more of an issue…….in fact, the filter inductor situation needs pretty massive damping to achieve the low zout of the input filter……and then you wonder if the damping of it is going to then permit the disturbances which afflict paralleled modules to actually reak their merry havoc.
    Last edited by treez; 3rd March 2020 at 08:55.



    •   AltAdvertisement

        
       

  17. #37
    Advanced Member level 5
    Points: 13,285, Level: 27

    Join Date
    Aug 2015
    Location
    Melbourne
    Posts
    2,180
    Helped
    795 / 795
    Points
    13,285
    Level
    27

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    We have dealt with a majority of Vicor modules - the internals ( i.e. mosfet switching ) have little to do with overall paralleling performance. The key to multi-paralleling is to ensure a low enough source Z - so that any one module cannot pull the input voltage around and affect other modules.
    2ndly, the cross-over freq ( I assume you mean where the closed loop gain = 1 ) also has little to do with paralleling - in fact a high X-over freq means that a small L will suffice to isolate the o/p from a large bus capacitance ( which you often get with a lot of converters in parallel ) - an L on the o/p of each means that the HF component of Vout can be seen by each controller of it's own power stage - i.e. it is not swamped by the large o/p C.
    3rdly, even though these modules have droop paralleling - additional small R always makes things better ( damping ) ....


    1 members found this post helpful.

    •   AltAdvertisement

        
       

  18. #38
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    Hi,
    The output overcurrent limit for the Vicor BCM4414xG0F4440yzz DCDC module is stated as 42A to 70A. Do you know what happens when this level is breached? The datasheet doesn’t say..
    We wonder if it latches off?
    ..Or maybe it turns off for a certain amount of time, then re-trys?
    ..or maybe it clamps its output current to 42A?

    BCM4414… module datasheet
    http://www.vicorpower.com/documents/...G0F4440yzz.pdf

    ............................*..................... ....................*............................. ............*..................................... .....
    ............................*..................... ....................*............................. ............*..................................... .....

    Hi,
    We are making an SMPS which will be at the end of a 300m cable.
    The Vin is 700VDC. This is from an isolated 700VDC source.
    Pout = 12Kw. (48vout)
    It will comprise “units” comprising of a vicor BCM module feeding four vicor DCM modules…..and these “units” paralleled in a whopping group of 10 to give the 48V, 12Kw output.
    We need to connect the chassis of the BCM modules to chassis ground or earth………also, the App notes state that under each DCM module there must be a copper plane, which is connected to either chassis ground or earth.
    The thing is, we don’t have an separate earth wire going to our product, and cant have this.
    We will simply be using 700VDC negative as our chassis ground.
    Also, we plan to connect the 700VDC negative to Earth…(so as to avoid this isolated output getting induced up to some high voltage).
    Do you think these measures are appropriate?

    ............................*..................... ....................*............................. ............*..................................... .....
    ............................*..................... ....................*............................. ............*..................................... .....

    Also, what must we connect our Y caps to if we only have a “go” and “return” going to the product?...and the chassis ground is our “return”
    …Or is it that we should simply just implement common mode EMC protection with a common mode choke only?
    The BCM and DCM modules are transformer isolated………so we believe to a considerable extent we can combat common mode noise by simply connecting primary and secondary with shorts……this will short out the transformer stray capacitances from primary to secondary we believe. Would you agree?
    Last edited by treez; 6th March 2020 at 09:01.



  19. #39
    Advanced Member level 5
    Points: 34,849, Level: 45
    Achievements:
    7 years registered

    Join Date
    Sep 2008
    Location
    cambridge
    Posts
    7,241
    Helped
    524 / 524
    Points
    34,849
    Level
    45

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    ............................*..................... ....................*............................. ............*..................................... .....
    ............................*..................... ....................*............................. ............*..................................... .....
    Hi,
    This concerns the DCM3623T50M53C2T00 DCDC module from vicor....

    Pg 23 of the DCM design guide from vicor states that the cut-off frequency of the input LC filter must be lower than the DCM module crossover frequency. (20kHz = crossover frequency for DCM module)
    However, the DCM3623T50M53C2T00 datasheet, on page 21, recommends an input LC filter of 1uH and 45uF (20uF plus the internal 25uF).
    45uF and 1uH gives 23.7kHz……this is above the DCM module crossover frequency…..do you know why they use an LC filter with a cut-off at a higher frequency than the crossover frequency?

    DCM3623T50M53C2T00 datasheet...
    http://www.vicorpower.com/documents/...53C2yzz_ds.pdf
    DCM design guide
    http://www.vicorpower.com/documents/...uide-VICOR.pdf


    ............................*..................... ....................*............................. ............*..................................... .....
    ............................*..................... ....................*............................. ............*..................................... .....
    Hi,
    This concerns the DCM3623T50M53C2T00 DCDC module from vicor.
    On page 26 of the DCM design guide, it states that for a “parallel damped LC input filter”, the Damping capacitor “Cd” must have a larger impedance than the LC filter capacitor “C(DM)” at the resonant frequency……but pg 26 then says that the capacitor “Cd” should be a larger capacitor than C(DM)………this doesn’t make sense.

    Do you have stipulations as to the blocking capactor "Cd"?....i believe it should be bigger in capacitance than "C(DM)"...what do you think?
    Pg 6 of the following ti.com app note states that the capacitor "Cd" should be 4 times the C(DM) capacitor, and from the calculation, this appears agreeable.
    http://www.ti.com/lit/an/snva538/snva538.pdf

    DCM3623T50M53C2T00 datasheet...
    http://www.vicorpower.com/documents/...53C2yzz_ds.pdf
    DCM design guide
    http://www.vicorpower.com/documents/...uide-VICOR.pdf
    Last edited by treez; 7th March 2020 at 23:19.



  20. #40
    Super Moderator
    Points: 54,106, Level: 56

    Join Date
    Apr 2011
    Location
    Minneapolis, Minnesota, USA
    Posts
    13,188
    Helped
    2628 / 2628
    Points
    54,106
    Level
    56

    Re: Large numbers of paralleled DCDC modules?...more than datasheet recomends

    This may not be cause for concern but it's known that atmospheric voltage increases with altitude. Articles such as Wikipedia state amplitude as 100V/m. At 300 meters this implies your drone is exposed to 30 kV. It seems incredible although it's static charge at pico-Amperes of current.

    Another cause of static charge could be air rushing around the craft (either wind or propeller blast).

    It's not necessarily a problem for your drone. It's not generally reported as a problem for planes/ copters/ ultralights/ model aircraft. However in a storm an aircraft risks being a lightning rod. Likewise a tethered craft. Or perhaps static charge could build to a point where you get sparks and endanger equipment.

    It's an indefinite part of your equation. I don't know whether you should drain the static to soil ground via spark arrestor, or extra wire, etc. I don't know whether it should be the positive or negative wire.



--[[ ]]--