VR1,2,3 getting very hot in less than 20s

[Solved] Re: VR1,2,3 getting very hot in less than 20s

Reply #15 – May 20, 2010, 08:15:22 pm

Hi,

Seeed sent me a new board, I just got my hands on it.

All Vr are outputing the right values, leds are on, seems to work. I'll try software later.

Thanks for your support : DP.com, GF.net and SS.com

Re: VR1,2,3 getting very hot in less than 20s

Reply #16 – May 21, 2010, 07:23:39 am

Good to hear that Seeed replaced your OLS. The problem with the "low 3.3V" and the hot VRs must (have) be(en) caused by some shortage on the board. I shorted 3.3V and GND one of my boards and toasted VR1 - totally my own fault! - and so I took the opportunity to repeat some of your tests.

Measuring the resistance between GND and the 3 voltage planes on two of my production pre-order OLS boards got me following results - pretty much consistant with what you measured (considering I used a cheap US$ 15 multimeter):

3.3V: 66kÎ©, 43kÎ©
2.5V: 19.4kÎ©, 20.7kÎ©
1.2V: 380Î©, 520Î©

I got my shorted board fixed by replacing VR1!

I noticed your question about the need for the 1.2v plane was never answered ...

[quote author="elbeem"]
... if the problem is around the 1.2 plane (why a 1.2 plane btw?), all the components could be implied, sic![/quote]

The technology Xilinx uses for their Spartan-3E FPGAs requires 1,2V for the core voltage VCCINT besides 2,5V and 3.3V. It's the silicon technology used in all Spartan-3 family devices (Spartan-3, Spartan-3E, Spartan-3A and Spartan-3AN) that determines the core voltage at 1.2V for all internal logic functions (as well as for the input signals for standards at 1.2V, 1.5V, and 1.8V). I hope this explains why 1.2V is required eventhough 1.2V, 1.5V and 1,8V input signals are not used in the OLS design.