Using power modules to get 175A at 1V5

[Flyback]

Sorry for the long and dull post...please by all means just skim read and hit back with anything you suspect may be remotely relevant.....

We are providing 1V5, 173A to a processor load by way of a 28V input.
This means using a cascaded network of Off-the-shelf Power modules by Vicor.com as in the JPEG attached here.
So, the 28Vin goes into a Vicor P024T048T12AL Power module, and this then feeds into a Vicor VTM48ET020T080A00 Power module which outputs the 173A at 1.5V (as in the attached JPEG)
(the power modules are actually paralleled to manage the large current throughput.)
The ‘upstream’ power module has a regulated Vout but the ‘downstream’ module is just a converter with a fixed Vout/Vin ratio (unregulated).

However, the ‘upstream’ power module purports to be able to do “Adaptive loop” control which means that the Load does not experience (much) droop of its supply voltage as it draws more current. (the upstream module adaptively increases its Vout as it supplies more current, and this is done in conjunction with user-set control resistors).

Question 1:
Do you know if this adaptive control actually still works when these power modules are paralleled?
..the following application note claims to offer advice on such adaptive vout control for systems with up to two VTM modules in series……why only two?....does adaptive vout control not work for more than 2 modules in parallel…
Application note on Adaptive vout regulation:
http://cdn.vicorpower.com/documents/application_notes/vichip_appnote24.pdf
Question 2:
Why is the MVTM36BT015M080A00 power module (datasheet immediately below) more expensive than the VTM48ET020T080A00 power module?...
..After all, the latter has a much lower output resistance.
MVTM36BT015M080A00 power module datasheet……
http://www.vicorpower.com/products?productType=cfg&productKey=MVTM36BT015M080A00
Question 3:
If we use the adaptive loop regulation with the P024T048T12AL/ VTM48ET020T080A00 Power modules, then given our load requirement, and supposing we have a 6 milliohm of trace resistance to the load from the output of the VTM48ET020T080A00 Power module, then what sort of load regulation could we expect? There is actually an application note which purports to offer a calculation for this (linked immediately below), however, this excel calculation spreadsheet will not allow me to adjust the cell contents…it asks for a password to do so, and I do not have this.)
App Note on Adaptive loop feedback…
http://cdn.vicorpower.com/documents/application_notes/vichip_appnote24.pdf
Question 4:
The bottom left of page 8 of the Vicor P024T048T12AL Power module datasheet states that…..
“Please consult Vicor Applications Engineering regarding additional considerations when paralleling more than two PRMs.”
..Why is this?...what’s the problem when connecting more than two P024T048T12AL Power module datasheets in parallel.?

Vicor P024T048T12AL Power module datasheet:
http://www.vicorpower.com/products?productType=cfg&productKey=P024T048T12AL

Vicor VTM48ET020T080A00 Power module datasheet:
http://www.vicorpower.com/products?productType=cfg&productKey=VTM48ET020T080A00

 

[bountyhunter]

“Please consult Vicor Applications Engineering regarding additional considerations when paralleling more than two PRMs.”
..Why is this?...what’s the problem when connecting more than two P024T048T12AL Power module datasheets in parallel.?

I suspect that is what they will explain to you when you consult them. Getting power converters to share in parallel is not trivial. I suspect whatever circuitry they have to do that has a limit as to how many other units it can drive (synchronize).

 

[spec]

I know you haven't asked this question but, I think you could simplify by using three 28V to 1.5V, 60A DC to DC converters in parallel.

SynQor have two units which appear to do the job: IQ24 and MCOT28. The nominal output voltage of the IQ24 is 1.8V but a -20% +10% trim is provided which spans 1.5V.

spec

Data Sheets
IQ24
**broken link removed**
MCOT28
http://www.synqor.com/milqor-milcots/mcots28.html
Three Parallel Units
**broken link removed**

 

[Flyback]

Beautiful!..Thankyou ...you read my mind. Thats a great contribution, will look into it now. I know of few manufacturers who do these kind of low vout modules.

 

[spec]

Beautiful!..Thankyou ...you read my mind. Thats a great contribution, will look into it now. I know of few manufacturers who do these kind of low vout modules.

No probs Flyback

spec

 

[Flyback]

Hello,
Regarding an array of paralleled vicor PRM modules feeding into an array of paralleled vicor VTM modules, how can we get good sharing of current through all modules?
Vin=28v, vout=1v5 at 175A

..Regarding such current sharing throughout an array to a load...isnt it a disadvantage to have this cascade of PRM modules going into VTM modules, then into the load?.....wouldnt it be easier to get good current sharing with a single stage 28vdc to 1v5?

 

[AnalogKid]

The answer to all questions about the ... personalities ...of Vicor products is exactly the same: Because it's Vicor. Their parts are weird, do not use conventional topologies, do not use conventional magnetics, do not use conventional current sharing schemes,,, and are weird. Their application support is excellent, but they rarely advise going more than about 1 mm outside the datasheets.

Synqor is a bit better, but not much. They use a digital current sharing scheme that can have start-up problems.

ak

 

[AnalogKid]

..Regarding such current sharing throughout an array to a load...isnt it a disadvantage to have this cascade of PRM modules going into VTM modules, then into the load?.....wouldnt it be easier to get good current sharing with a single stage 28vdc to 1v5?

It would, but the Vicor Factorized Power system was developed for a specific market and group of applications, then generalized for a broader market. At its heart it is a point-of-load regulator like an overgrown 7805, not an industrial power system. It is by far the strangest high density power conversion system out there. Vicor's newest parts are trending toward more conventional configurations, but I don't see any 1.5 V output parts yet.

ak

 

[Flyback]

Thanks, we originally thought that the vicor VTM modules that convert the 36V to the 1v5 would not share current with each other when paralleled, but in fact they will……they will just share it in proportion to the PCB trace resistance that exists between each paralleled vicor module and the (common) load.

As far as I can see, the vicor method of having a cascade of PRM modules feeding into VTM modules is a way of getting it so that you don’t actually need much capacitance at the load….which means no electrolytics at the load so you can make the circuitry lower profile…the tall electrolytics can be placed further back from the load, out of the way.

The Vicor VTM modules that convert the 36v to the 1v5 have the property of “capacitance multiplication”, whereby they actually effectively multiply the capacitance value at their input and make it seem that that capacitance is at its output…effectively, -thus you don’t need the big caps at the load…..that seems to me to be the modus operandi of the vicor situation where they have the PRM SMPS feeding into the VTM SMPS?