|
Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets. |
|
Thread Tools |
13th Sep 2019, 8:50 am | #1 |
Dekatron
Join Date: Jul 2008
Location: Kington, Herefordshire, UK.
Posts: 3,658
|
Capacitor temperature coefficients
Bit of a mouthful, but the query is this. I'm working on a bit of 60s test gear which has a multivibrator putting out a nominal 3150 Hz. This uses two polystyrene capacitors (5nF) and an NTC resistor to compensate for their drift with temperature, and there's a fair amount of that, despite being transistorised, as there's a lot of R/C smoothing in the power supply. Simply placing fingers on the caps causes the oscillator to drift up by several Hz, and over several hours it drifts by about 30 Hz, so the NTC isn't tracking the caps too well.
Now, without doing a wholesale rebuild of the stage, it seems to me that careful selection of capacitor type might give me something which reduces the drift substantially, either by bringing it closer to the curve of the NTC, or being sufficiently tight that I could dispense with it. Any ideas? |
13th Sep 2019, 3:29 pm | #2 |
Dekatron
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,953
|
Re: Capacitor temperature coefficients
Not sure about R-C oscillators, but in the world of RF inductor/capacitor variable-frequency-oscillators it's almost universal to include a degree of negative-temperature-coefficient in the capacitor(s) used to resonate the coil. These are usually ceramic "N750" capacitors, usually a few tens/hundreds of pF.
You may be able to improve your oscillator by replacing the 5nF polyesters by a smaller value of polyester paralleled by some N750 ceramics - but the process would be one of trial and error to find just what ratio of polyester to ceramic would give the right result. |
17th Sep 2019, 10:42 pm | #3 |
Octode
Join Date: Dec 2008
Location: Wincanton, Somerset, UK.
Posts: 1,757
|
Re: Capacitor temperature coefficients
It's difficult to comment with any certainty without seeing the circuit, but I wouldn't expect polystyrene's to cause significant temperature drift: -150ppm from memory. So 10 degrees would be less than 3Hz at 3.1kHz.
More likely the delta Vbe of the switching transistors (or diodes?) causing the multivibrator to change time interval? I saw this on a Tek 422 calibrator which had appalling frequency drift (due to effectively to 2 Vbe's/diodes in series). I modified the circuit to eliminate this problem. |