Op amp circuit problem

[1uk3]

Hi

I have built an amplifier for small and short current pulses. The second stage is CR-RC²-shaper. It works quite well but i have noticed some (for me) weird behavior


The Problem:
As soon as i connect T1 ( 1.5m RG59/U  75?), i get a -2V offset at the output. The other end of the cable is open! Terminating it with 75? doesn't change anything
I can fix it by connecting a scope probe at TP1 and the offset disappears, even if the probe is not connected to the scope.
The other weird thing is that the inverting input of U1 is also slightly negative (should be 0V right?)

All op amps are LMH6682 (datasheet http://www.farnell.com/datasheets/51551.pdf)
D1 and D2: BAV99

What I've tried:
First i thought it is a bad solder joint, and the force of probe helps to make good contact.
Then i replaced the probe with a 1M? and a few pF. Didn't work. I tried different values and either get no effect or oscillations.
The supply rails are stable. Checked each solder joint a million times.

When connecting an "antenna" i would have expected noise or 50Hz at the output, but a DC offset?

I have no idea what the else i could do, would be very grateful for every advice.

Thank you!

[dom0]

- Photo of your build
- High speed amplifiers can oscillate with low amplitudes that might be hard to see when checking the output at several V/div, but even these low level oscillations can disrupt operation.

I don't really understand what you are trying to do with the open cable on the input? Can you explain your application a bit more?

[1uk3]

Thanks for your reply 

Can't give you a photo because I'm not at home for a week now.
What do you mean with small amplitudes? <10mV? I couldn't find anything with my scope (ac coupled 10mV/div)

On the other end of the cable is a high impedance sensor, but the problems started as soon as i connected the cable.
I want to measure the area of a short current pulse with a relatively slow ADC.
The first amplifier integrates over the peak, the following stages stretch the pulse.
V(n006) is the output.

Yes i know that there are precision peak detector circuits but i want to try it this way.

[1uk3]

dom0: You were 100% right 

Very stable oscillation at >300MHz. I can only guess frequency and amplitude because i only have a DS1054Z (100MHz, 1GSa/s)

Works perfectly fine now
Thank you very much!