Monday, July 19, 2021

Non-inuitive aspects of getting Solar

What I didn't expect before I got solar panels on my house:

  1. The wattage on the Solar Panel does not directly correspond to the actual output. 
    1. Your actual maximum generation will be determined by your inverter.
    2. The output of the solar panel is not constant and on cloudy days the production can swing by over 8x over 15-30 minutes.
  2. The size of the inverter is likely going to be less than the combined CEC number for all your panels and that's probably ok. Losing some high production days to 'clipping' may be better because the higher wattage panel will increase production around that point.
  3. NREL has a lot of useful tools. pvwatts and SAM
  4.  If you get snow, you want to look into snowguards for the solar panels
  5. The orientation of the panel for the snowguards may (or may not) matter - look at how your panel handles shading.
  6. Solar without batteries currently doesn't provide you any power in the case the utility grid goes out.
  7. When the power goes out and you are running off the grid, too much production from solar can be as bad as too little. 
  8. The Tesla powerwall, when off grid and needing to curtail the production of the inverters will adjust the frequency of the electrical system. This adjustment may trip your UPS and make it go active.
  9. The Tesla powerwall system (i.e. the  Tesla Gateway) does not guarantee fast switching so you'll still need a UPS for anything electrically sensitive.
  10. If you have Time of Use (TOU) rates be sure to know when your peak rate buy-back time is. Conventional wisdom is to have your panels facing south for the greatest production but it may be more cost effective to have westerly facing panels since TOU rates tend to be higher in the late afternoon and evening.

Monday, May 03, 2021

Blue light glasses and video calls

While obvious in hindsight.... blue light blocking glasses means that you're going to have big blue blocks reflecting from your glasses when you're on a video conference from your desktop/laptop. 

 Turning on "night light" mode seems to help a bit by reducing the blue output from the screen itself.

Monday, April 06, 2020

Dell R740xd2 PCIe x8 x4 slots

Did you just get a brand new Dell R740xd2 and noticed that for some odd reason Dell didn't pre-install an PCIe x8 card?

Did you then go looking around for a PCIe x8 slot and find everyone one already taken up?

Well, what you may have missed was that slot 5 is actually an open slot where you can put in the x8 card...

This is why the Dell docs list slot 5 as an option for a variety of cards:
https://www.dell.com/support/manuals/us/en/04/poweredge-r740xd2/per740xd2_ism_pub/expansion-card-installation-guidelines?guid=guid-b1d52a5e-f568-4a7a-a2cc-eb5ddb6dedc5&lang=en-us

Page 20 of https://i.dell.com/sites/csdocuments/Product_Docs/en/dell-emc-poweredge-r740xd2-technical-guide.pdf also has the following table for PCIe mappings:


Wednesday, May 24, 2017

Skull Canyon NUC UEFI install

I was trying to do a clean install of Windows 10 on the Skull Canyon Intel NUC (NUC6i7KYK).

At first, I thought there was something goofy with the USB stick that I was using as the UEFI boot (F10) would see the stick but when I selected it as a boot device, it would just return me to the boot menu and apparently do nothing.  Tried different USB stick; tried different ways of building the USB stick; tried different USB ports; tried to update the BIOS (from 0041 to 0046).  No joy.

What ended up working was enabling the UEFI shell from the BIOS.  After booting into the UEFI shell, I could then specify the USB stick and then manually run \efi\boot\bootx64.efi.  Install worked fine after that.

Thursday, November 27, 2014

Basis Peak vs. Carbon Steel

I just upgraded from the Basis Carbon Steel to the new Peak.  The data was preserved in the migration and all I needed to do was install a new app on my iPhone and log in again.

Some random thoughts:

  • The screen is larger but it is less thick. As such it feels smaller in some sense.
  • The new charger doesn't require the unit to be slid in; you can just drop it on there. However, it seems to have the same problem that if you aren't paying attention, you can put it on the  'wrong way' and even though it is seated snugly, it doesn't appear to charge.
  • A 'power brick' for the USB adapter wasn't provided. I don't mind since I have the Anker 40W 5-port USB charger.
  • The gestures you use to do things (turn on/off backlight) isn't obvious and the documentation is hidden at http://support.mybasis.com/hc/en-us/articles/203483960-Display-Modes.  I would have thought there would be a quick cheat-sheet in the packaging but I don't remember seeing it.
  • The band feels more comfortable. I don't know why.
  • Firmware updates can be done from the phone now so it doesn't look like you need to periodically attach it to a computer.
  • Sync'ing is now continuous so you don't need to remember to manually sync (or forget and fill the memory of the unit).
  • The constant connection to the iPhone 5 doesn't seem to be a bad battery drain
From a marketing perspective for current customers, they offered a $30 credit after I had already pre-ordered and received the Peak from Amazon.  That was really bad timing and what is disappointing is that they aren't offering that credit towards something like a second charger. 

At this point, none of the advanced functionality (seeing texts on the screen, stop watch, etc.) is available which is a little disappointing. Hopefully, it won't be too long before these updates come out.

Sunday, July 06, 2014

ServiceNow: gs.dateDiff() returns the wrong results (sorta)

Do you see what is wrong with this code?
var date1 = new GlideDateTime(); 
var date2 = new GlideDateTime(); 
var date3 = new GlideDateTime();
date1.setDisplayValueInternal('2011-08-26 17:12:28'); 
date2.setDisplayValueInternal('2011-08-30 12:33:10'); 
date3.setDisplayValueInternal('2011-08-30 12:43:10'); 
gs.print(gs.dateDiff(date1, date2, true)); 
gs.print(gs.dateDiff(date1, date3, true));
If you run the code, you'll see that gs.dateDiff() returns the same value for the difference between date1/date2 and date1/date3 even though there is a 10 minute difference.  Why is it rounding to the full number of days?  

Well, it turns out that the default string conversion is not what gs.dateDiff() is expecting.  Instead, it is expecting the date/time string in the user's display format.  

(so if your default date / timeformat matches the internal, then you'd actually get the correct results and wonder what I'm complaining about.) 

Instead of just failing, gs.dateDiff() seems to do the math where it can (e.g. the difference in days). This seems a poor choice in that you are giving the impression of valid results when the results are not accurate.

What you need to do to make sure you get the correct results is the following:

gs.print(gs.dateDiff(date1.getDisplayValue(), date2.getDisplayValue(), true));
gs.print(gs.dateDiff(date1.getDisplayValue(), date3.getDisplayValue(), true));
getDisplayValue() converts the internal date format to the one expected in the execution context.





Saturday, July 05, 2014

Servicenow: obvious in hindsight

Another 'duh' moment.

The use of gs.addInfoMessage() will not work with business rules set as 'async'. If you want the message, you need to wait and use 'after'.

Yeah, this is obvious in hindsight as the execution of the BR may not immediately occur so you can't get the message to appear on the next loaded page.