Topics

Sampling with 7S14


 

Hi all sampling gurus,

I am continuing performance check of my 7S14 (B020274). Any idea how many pieces were manufactured?
All the Hg cells were empty (and were leaked also). I replicated with the concept from Miguel.
Some photos uploaded to "Sampling with 7S14" album.

At least ch1 seems to be working properly. Quite noisy I thought (picture labelled first try).
Ch2 shows some sporadic operation.

I measured the power supplies and noticed that +30V was over 40V and -5V was -5.3V.
Replaced faulty U1 and U2 which fixed this problem. Just wondering if this has caused some damage?
I didn't notice tantalum caps issues (didn't remove any boards).
Some push button exercise with IPA and lubricant.

After those actions everything seems to be operational (all other photos).
Noise level seems to be lowered. I didn't dig into schematic but could those wrong voltage levels increase the noise level?

I haven't used 7S14 in my career. Does the screen images look ok?

Best Regards,
Jouko


Ed Breya
 

Did you include the bypass capacitors on the photo-voltaic bias outputs? There should be one across each, so four altogether. They are essential for best noise performance. Ed


 

Hi Ed,

1uF ceramic cap in each.

BR,
Jouko


Carl Hallberg
 

Ed,
Wow.
How we keep beating this to death.  I gave data on different LEDs,  gave you credit for suggesting capacitors across outputs and I used 15uF because I have lots of them, by experiment used 3W red LEDs because of higher outputs to supply enough current,  another person did SPICE evaluations on Red with wavelength recommendations.  Another mentioned that exact 1.3V wasn't important whereas my 7S14 was approximate 1.5Volts.  I gave photos for the Tek group.  


Carl Hallberg

On Sunday, February 28, 2021, 01:02:12 PM CST, Ed Breya via groups.io <edbreya=yahoo.com@groups.io> wrote:





Did you include the bypass capacitors on the photo-voltaic bias outputs? There should be one across each, so four altogether. They are essential for best noise performance.  Ed


Ed Breya
 

Sounds good. Ed


Ed Breya
 

Carl wrote:

"Ed,
Wow.
How we keep beating this to death. I gave data on different LEDs, gave you credit for suggesting capacitors across outputs and I used 15uF because I have lots of them, by experiment used 3W red LEDs because of higher outputs to supply enough current, another person did SPICE evaluations on Red with wavelength recommendations. Another mentioned that exact 1.3V wasn't important whereas my 7S14 was approximate 1.5Volts. I gave photos for the Tek group"


Carl, you seem upset about something, and I don't see why. I'm not "beating this to death" that I know of. I saw the post about possible noise issues, and suggested to be sure the caps were included. I don't keep track of the various implementations done over the years, but if the caps get overlooked, it can be a problem. I don't know anything about your version of it, except what you just mentioned.

Ed


Roger Evans
 

The noise level displayed on sampling systems is sensitive to the setting of the loop gain for the DC balance amplifier (R231 and R232 on the 7S14). If you change the sampling diode bias voltage then you should set the loop gain accordingly, but this is all part of the calibration anyway and is well described in the manual. Basically you look at a fast edge at a relatively low sweep speed and adjust for no undershoot or overshoot, the noise increases rapidly once you get into the overshoot region.

Roger


 

Hi Roger,

I tried to tweak those loop gains. They seemed to be in optimal setting.

I am satisfied with the performance now (photos from 28th Feb in the album). I think that the issues with the power supply created the noise in my first try (photo from 26th Feb).
Or there is some intermittent fault still lurking somewhere...

Best Regards,
Jouko