Thanks for the rom and your detailed explanation. It looks like you are pretty much through the whole bios modding process as far as MBT goes. I thought you maybe mixed up the voltage table but after reading your text above I guess you understand how the voltage table works, which was the only point I wanted to take a look in your bios.
I will look into it anyway but I suspect its more and more a KPE specific behavior like not showing the right voltage above 1212 mV in tools. I know that the max. value for standard cards is 1281 mV. All values above they can’t recognize and the bios is acting weird with such settings. This leads me to the question, because we know that the Ti cards always put one voltage step more then set. Have you tried setting the top three sliders at 1275, does windows still won’t boot? I’m now confused myself because in my thread I wrote that the max values are 1281 mV even for the Ti cards and after all these years there wasn’t anyone pointing out that his 980 Ti won’t boot after setting it to 1281 mV. So a possible try would be to limit your top 3 sliders to 1275 and try booting windows again.
When I’m home (now I’m at work, stupid timezones) I will take a look into my old bios collection and if there is no ti bios with 1281 on the top sliders then you have found a 3 year old mistake which nobody ever noticed 
.
Regarding your card boosting to clk 61 instead of 52, was that with or without the +25 mV switches at the back of the card? Also the voltage table isn’t that precise. It does not mean that you can’t access a clk state which does have a higher value on the right side then your top slider is set to. The left value of the clk state does have a role too. That’s what I hate about nvidias gpu boost. You just can’t tell exactly at which clk state the card stops only by looking at the voltages, because of the voltage range each clk state has. And thats the reason you could boost up to clk 61 even the right values where higher (and the left ones were lower) than your stock bios voltage settings. But for full voltage table access you need to edit the right and left side like on the picture to make the card boost up to clk 74. I hope that’s somewhat understandable.
And yes the maxwell wall is defiantly starting at 1500 and being really hard at 1550. Even more sadder is that also a KPE card does suffer from that wall. I mean whats the point of the KPE card when every good standard 980 Ti can do 1510 ± 20 MHZ too 
No problem. Someone on EVGA’s forum told me that the Kingpins voltage is locked down to 1212mv (+50mv if use dip switches) and that only way to raise the voltage and get real time monitoring of the voltage is either through Precision X with Overboost and voltage control enabled or use EVBot module. There might be a third way. I was able to verify that using Precision the voltage was in fact raised and accurately reported shy of 6mv but unfortunately that voltage was locked and wouldn’t down volt under idle conditions.
Trying to set a higher than 1212mv voltage by unlocking the three voltage control tables to 1281mv using Kepler Bios Tool caused windows to not complete booting into the desktop. Read that was an intentional driver fault from NVidia though I can’t be sure what was going on to cause Windows to fail booting all the way. I should have taken a volt meter to the cards and checked but didn’t think of it.
About the boost state going to CLK #61 that was with and without. I’m like 90% sure because I turned off the voltage boost and it still was boosting to 1430 at Clk state #61. Anyhow I think your wrong about the right side of the table controlling weather or not the CLK state will be accessed or not because I looked at my tables and certain clock states still got boosted into despite the max exceding the voltage map max value. I read an article about Kepler boost states last week that stated nvidia has some algorithm it uses for that but can tell you if both CLK values are over the max map value it definitely will not be used. Sure we can both agree on that.
Anyhow it was disappointing that despite the much cooler temps that my Kingpin cards cant go over like 1510 mhz despite voltage raises. I mean I was loading at 32 degrees last night with Furmark! Maybe because I got them in SLI I’m limited. Maybe because my ASIC scores are 75.0 and 76.5. Though found out after the fact that the Kingpin cards are only meant for LN2 in regards to overclocking despite the custom PCB and extra power phases and VRMs. Putting them on water doesn’t help.
Going to do more experimenting with overclocking them further later this week. Got an Evga Probe-It cable on the way. I personally think that the Boost states up and down are what cause instability at high mhz. Much like when your old school over clocking your cpu its best to disable EIST and lock in your frequency at high volts as the adaptive overclocking methods are so much more difficult though not impossible. So got more testing to do.
Thanks