Good work. I love to restore fine old piece of equipment with attention not just to get them working but really restoring the original design.
On Sat, Jan 7, 2012 at 3:17 PM, Frank Stellmach <[email protected]>wrote: > Dear time-nuts, > > the 10811, sitting in my HP5370B, showed an intermittent heater circuitry > failure. > It turned out, that the NTC temperature sensor was damaged (1982 vintage). > > At room temperature it was conducting (100kOhm), but became open during > heat up phase, eventually conducting properly at around 82°C, but failing > from time to time, i.e. the star symbol on the 5370B turned on frequently, > and the reference frequency fluctuated strongly. > > Normally this is a total damage of this OCXO, as the NTC is epoxied > completely into the oven mass, and this whole assembly would have been to > be replaced, according to the manual. > I doubt, that it's available any more, but definitely not for a reasonable > price. > > I found a compatible NTC replacement, regarding diameter (2,41mm max) and > electrical parameters, i.e. Rn(25°C) = 100kOhm, B = 4540K, and an even > better accuracy of 1%, i.e. 8.6kOhm at 82.8°C +/- < 1°C. > It's from epcos, S57867S104F140. Visit epcos.com for data sheet and R(T) > table. > > This NTC has 50mm short and blank wires; I couldn't get the PTFE insulated > S861.. type. > So the wires had to be insulated with a heat shrink tube, and the NTCs > wires had to be assembled straight, without bending them 180 degrees > through the 2nd hole in the oven mass. That's also much better regarding > reliability. > > After complete disassembly and carefully milling out the old NTC and the > epoxy with a 2.5mm drill, I applied thermal grease at the new NTCs head, > and fixed the wires with a small drop of epoxy at the other side of the > oven mass. > > The reassembly was a little bit delicate, due to the short wires, and > because the insulation of the wires was thicker than before. But finally, > the heater circuitry stabilized normally within 10 minutes. Obviously, the > thermal gain was not perceptibly affected by this slightly different > assembly. > > The OCXO now works fine since two weeks, and its drift is in the order of > < 1e-9 again. > > Frank > > > > > > _______________________________________________ > time-nuts mailing list -- [email protected] > To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
