Mercury getting into 500 series

Very interesting.....though I wonder if the Mercury modules will require power only available in their own rack....

+1

Except no one makes 24V modules really, except DIY guys, and they kinda gave up on the format after the rack makers started releasing racks with more amps per rail.  Voltage Converters are handling the power needs of modules that relied on the 24v rail now.  Just ask Mike from Serpent Audio :-( 

I take u at ur word Brian, I'm know I'm conditioned, but i would never hot swap a unit myself, what's ur take on the 2.5 amp thing? I'm trying to wrap my head around that price tag, It just seems über expensive to me

You know, as a designer, having 250mA per rail per slot rather than the 125mA in the VPR spec is certainly welcome.  The problem I have with it is purely from a marketing standpoint.  Your average user who doesn't know an amp from a dB won't immediately understand that some modules should only ever be installed in these newer racks.  I wish there was some industry-wide "500 v2" or some such thing that could be branded on both the modules and the rack showing that a high-draw module will only go in a 250mA per slot rack.

Imagine the confusion when someone who doesn't understand these things at all puts four 250mA cards in their old lunchbox and then when they put in a 5th, the power supply craps out.  There will be much cursing and finger pointing.

I think most people know the old racks suck.  I barely see them anywhere any more.  As a designer, I would design for the future not the past.  

I now know a number of guys using the soft start in there modules to control the in rush.  Works great, and is really smart.

My question is, do we really need another 500 series anything?  We have such amazing tools as is.  I hope they don't just bring another version of everything else.  Something new and creative would be appreciated.  

They mention individual regulation per slot.  They also mention the separate PSU being linear, so it's probably what we call an "intermediate power bus" with a higher voltage and then the local step-down regulation at the slot is the "point of load regulation".  It's pretty common in complex designs to have some kind of bulk high voltage bus and then regulate as close to the load as possible.  That's so changes in the load are regulated immediately due to the close proximity.  If the regulators are 3ft away in another box and that step in current has to travel that far through thin/long conductors, you're going to have some distortion depending on the regulation and decoupling in the 500 card demanding that current change.  If you regulate locally then you drastically cut down the time needed to regulate that step in current.

That's the theory anyway.  


I use it in my high speed designs because the ICs I use can draw Amps of current instantaneously.  Since 500 series specs are much less than that, it might not be such a big deal.  It'll really come down to the 500 device you plug in and the amount of decoupling it has to supply the instantaneous current requirements.

EDIT:

I'm not sure inrush is a big issue on 500 series modules, it would be a bigger deal for the rack power supply though.  Imagine that the supply can increase the voltage to account for the larger current draw, but it does so slowly.  Since V=IR and the load, R, is constant, then V and I must change in an inverse manner.  So as your module is drawing larger currents (inrush) to charge all of the caps, traces and ICs on the board, the voltage will drop.  The power supply will increase the voltage in an attempt to regulate.  The speed of this is known as the "step response".  If it's too slow then you'll see your current drop and the power supply will jack the voltage way up, but as the current draw tapers off and settles, then the voltage will stay too high until the supply can bring it back down.  This can be a volt or 10, all depending on the supply.  10 extra volts might kill certain devices, or it might not.  Again, we'd be back to how the module was designed.  I'm assuming that they use linear regulators for local regulation, which would be pretty fast.  Most of the time the step response problem I mentioned happens in poorly designed switch mode supplies.


Now, if you were to have hot swapping, you'd need to control power timing so that you don't spark/arc as you plug the card in, but with such low currents I don't see a big inrush problem here.

I think what I'm saying is that this is mostly marketing.  Local regulation is a good thing, but I'm not sure they are solving a widespread problem here.  It's like Low-T treatments you see on TV commercials, a solution looking for a problem to solve.

Yeah, inrush currents can pop fuses or trip breakers for sure. It could mean that the 660 has a lot of parts to charge up, or it could mean that the rack uses power traces that are much too narrow/thin which acts like a resistor under higher currents, which can cause excessive current draw. The rack probably uses something called a PTC, A.K.A. re-settable fuses. The first turn on charges up the devices and trips the fuse, then the second power up doesn't because the device's caps are already formed and partially charged.

John, buy a Splice-500 from ruckus. it'll smoke any other 500-series 1176 clone out there, hands down.

oh lol

These available in kit form, Randy?