Programmable Devices
CPLDs, FPGAs, SoC FPGAs, Configuration, and Transceivers
20641 Discussions

Powering Stratix IV Dev kit via HSMC

Altera_Forum
Honored Contributor II
1,049 Views

Hi, 

 

Has anybody had any experience powering the Stratix IV dev kit by powering the 3.3 and 12V lines on the HSMC connector instead of the normal DC input/PCI express connector? 

 

The ref manual doesn't support it. But I don't see any reason why it shouldn't apart from power measurement sensing circuit connected to the system controller via I2C. 

 

Thanks 

ZubairLK
0 Kudos
4 Replies
Altera_Forum
Honored Contributor II
266 Views

The rails you refer to on the HSMC connector are, typically, there to provide power to any daughter board connected to it. However, the 12V rail is usually simply routed there from whatever primary source is powering the development kit - perhaps via a protection diode. 

 

You don't specifically state which Stratix IV development kit you are using but a good look at the schematics should reveal how the 12V gets to the HSMC connector on that particular board. If it is routed directly and you're not supplying power from another source via another route onto the board, then it will be fine to supply 12V via the HSMC connector. 

 

The 3.3V rail is likely to be different. On all the development boards I'm aware of, that have HSMC connectors, the 3.3V rail is generated on the development kit. You should not attempt to supply 3.3V to the development board, via the HSMC connector, if this is the case for your kit. Doing so will clearly result in unwanted contention and could damage the dev kit or your board. 

 

If you identify which kit you're using I can have a quick look over the schematics to confirm whether this is the case. 

 

Regards, 

Alex
0 Kudos
Altera_Forum
Honored Contributor II
266 Views

Sorry. My mistake. 

 

I am using Stratix IV GX FPGA Development Kit 

 

The 230 edition. 

 

http://www.altera.co.uk/products/devkits/altera/kit-siv-gx.html 

 

In the reference manual, page 2-63 (pdf page 71) shows the 12V and 3.3V rails are directly connected to the hsmc. 

 

I checked the schematic as well. Everything seems ok to me except for 1 thing. 

 

The power measurement circuitry uses 12V_out (from the DC jack) and 12V to HSMC 

(Labels from schematic. Page 21). 

 

And these are connected to LTC4151. A power measurement IC. (ADC with a i2c interface).  

page 25 of schematic. 

 

The i2c interface goes to the EPM2210 System Controller.  

 

I don't know whether the system controller would trigger an auto-shutdown when/if detecting current flowing in the opposite direction.. 

 

Thanks 

ZubairLK
0 Kudos
Altera_Forum
Honored Contributor II
266 Views

Right - I'll back-track a little... 

 

For this kit both 12V & 3.3V are generated locally or supplied 'directly' via PCIe. So, forget my previous comments about treating 3.3V differently to 12V. If you're going to power this board via the HSMC you will have to supply 12V & 3.3V. 

 

Will the board be happy with this? In short, I think so. 

 

The LTC4151 current monitor might see a 'negative' current through R64. However, given the only loads on 12V_OUT are U15 & U19, the current drawn will be tiny, around 1mA. Across that sense resistor that doesn't amount to enough to register a current in the LTC4151, assuming it can report negative currents at all (I don't think it can). 

 

I don't know, but I'd be very surprised if the MAX controller was designed to shut the board down under these conditions. High currents - OK, maybe. 

 

When the board is powered via the PCIe connector (with U35 switched off via power switch SW1) U15 is active and U19 is not. In the same way D45 becomes active (for 3.3V) and D44 is not. Both LTC4357 & LTC4352 are designed to work in a system where they are off permanently. When off they ensure power isn't fed back through their respective FETs. 

 

Connecting power in the way you want will see all four of those parts remaining off (due to lack of power from either U35 or the PCIe connector), which in turn will hold their respective FETs off, preventing any back powering of the circuitry on page 21. 

 

So, I think (disclaimer: read that again - i think) powering the board via the HSMC will be fine and you'll be unlucky if the MAX system controller shuts the system down. If it does, I suspect we could work around that... 

 

Let me know how you get on. 

 

Cheers, 

Alex
0 Kudos
Altera_Forum
Honored Contributor II
266 Views

 

--- Quote Start ---  

 

The LTC4151 current monitor might see a 'negative' current through R64. However, given the only loads on 12V_OUT are U15 & U19, the current drawn will be tiny, around 1mA. Across that sense resistor that doesn't amount to enough to register a current in the LTC4151, assuming it can report negative currents at all (I don't think it can). 

 

--- Quote End ---  

 

 

Makes sense. Thanks for the input! 

 

I'll update when I check if it practically works. 

 

Cheers 

ZubairLK
0 Kudos
Reply