On 9/11/22 18:41, Tom Lee wrote:

Linear vs. SMPS would not be expected to matter, in any case. The problem is one of moisture ingress. The encapsulant shrinks and cracks over time, letting moisture in. At some point the capacitor breaks down and you get a multimedia spectacular. None of those processes is particularly sensitive to what circuitry lies downstream.

? It's my understanding that it doesn't happen quite like this. My understanding is that the epoxy is hygroscopic, and its mass increases as it absorbs moisture.? The mass of epoxy swells, causing first crazing, then full cracking of the outer shell. Leakage increases and the resultant current flow causes heating, until it goes bang.? I believe the end stage there is thermal runaway.

??????????? -Dave

Hi Bruce,

It would be interesting to see the quality of the scans.

Sample scans in the files section
A temporary directory for photographs and help relating to emails and posting ( /g/HP-Agilent-Keysight-equipment/files/A%20temporary%20directory%20for%20photographs%20and%20help%20relating%20to%20emails%20and%20posting )

... the 8510B (both the 85102B, and the 85101B) If anything.

Nothing on these - it looks like the instrument was introduced in the 1985 catalogue and so is outside the date range of the fiches that I have - 1969 - 1979 (-ish)

Cheers

Peter

Hi All,

I spoke to the fiche scanning house today - after review of the scanning process for a set of sample fiches that I sent them (HP 5480A) they were able to more accurately price any further job.

The fiches that I sent were typical of the earlier samples in the collection (some clean others quite scratched) and so required some post processing to separate the pages which put the per fiche price up a bit.
Their quoted cost is 1.07 UK pounds for each 14 x 7 fiche.? This works out at around 1.1 UK pence per page with a small discount likely for quantity

Would it be of use for me to upload some of the resulting .pdf files for people to get an idea of quality?? 100 pages runs to around 3MB

Peter

Hi there again!

Today a new problem child arrives at home, ‘79 8660C sig gen with 86634A phase modulator, 11661B freq xtender and 86603A RF section. Not a big fan of this HP era, but a few bucks for 33 kg of seventies stuff… nobody becomes bitter by a candy.
Here you have my first impressions:

Cosmetic issues, some rust in the aluminium frame, dust inside, two keys broken (the stud still in, so easy repairable), attenuator knob in RF section needs glue too, so, not a big deal on that side.

When I opened the pannels, saw a fantastic Rifa capacitor exploded, inmediately extirpate. Then I’ve started the unit using variac, it slowly turns on without smoke or explosions. Keys, switches, RPG, anunciators, led display and meters roughly working at first sight in both mainframe and plugins. Let’s try to see something at scope. Power seems ok, a bit over the numbers I read on meter, but not bad. *Frequency is 5.6 MHz over the selection on the display* and keep this behavior. You can see this in my SA screen. Obv it doubles the error at 2 GHz as you can see in my counter. Also the spectra is wobblin’ a lot with 100 kHz of RBW, with less resolution the skirt isn’t a skirt, only high noise. Terrible.

Obviously have a loop problem. But first things first. Power supply rails are quite spot on in dc measurement, *but ripple is terrible.* Figures are three and four times?over the specs. Also I’ve cheked unregulated rails for ripple and two of them have 250 and 300 mV over the specs. In the scope (AC coupled) the waveforms look weird, like a ramp or triangle. I think some caps are dry and maybe some diodes are blown in bridges. Besides this, I saw some corrosion and blackened area in the A20 board, near C4 and C5 capacitors. Tomorrow I’ll take appart the caps and boards to evaluate them.

During measurements I noticed that the corner of the instrument that holds the power supply guts are quite hot. Actually the whole machine was, but power related parts the most. Main transformer itself reaches 50?C (122?F) and the rectifier bridges almost 55?C (131?F). I think that isn’t normal. I’ll try an experiment: leave the unit cool down and connect it again but without turn the main switch on, thus if a capacitor is drawing current, I’ll be noticed with some temperature increment. Anyway, I’ll disassemble all parts to check them properly.

Some work to be done. Again, opened ears to every chap that wants to contribute with an experience or tip. I’ll be posting updates of my adventure. Mainframe op & service manual was easy to find, but nothing about the other three parts.

Regards,
Ruben

Hello Group -
My whole life I have been calibrating VNAs and then verifying the cal by connecting Through, and looking for very good RL and IL across the band.

I'll set up my confusion / question this way: let's say I have two VNAs that are perfectly calibrated. Now I take port 1 of VNA A and connect it to port 1 of VNA B.
What I am doing is measuring VNA B's actual impedance on port 1. Which is not going to be perfect - or necessarily close to it.
The "true" impedance of any VNA port is not as good as you measure after a cal with a through.
Why is that?

After a perfect cal, with a perfect through, S11 and S22 should be showing the actual "true" impedance of the other port (port 1 is looking at Port 2, and vice-versa.

Can someone tell me what is wrong with my thinking on this? I must be wrong... but I don't know why.

Thanks in advance -
Dan