Pocket Oscillator

[ruffrecords]

My pocket oscillator design used a pair of diodes as a non-linear resistance to ensure the overall gain of the Wein bridge oscillator stays close to 3 times for minimum distortion. Unfortunately it does not work too well even though I included a trim pot so you could tweak it for minimum distortion. I know that "classic" designs did not use diodes but instead used a varistor or thermistor (not sure which). I can find very little online about how to use these in this application. Anyone know how I go about selecting the right one?

Cheers

Ian

[mediatechnology]

Ian -

https://www.nutsvolts.com/magazine/arti ... ook-part-3

Looks like the RA53 is the go-to part number: https://www.farnell.com/datasheets/66949.pdf

Looking at the I-V curve may provide a clue.

Wayne

[ruffrecords]

Thanks Wayne, that is exactly the answer I needed. You got it in one. I did loads of online searches but never came across that article. It even shows the exact configuration I am using with back to back diodes and notes that the output distortion will be in the 1% to 2% region which is exactly what I get. I will get myself and RA53 an try it out.

Cheers

ian

[ruffrecords]

Looks like the RA53 is made of unobtanium so I searched online for thermistors and now I am confused again. All the thermistors I found were either intended for temperature measurement or for inrush current protection in power supplies. I am not even sure if they are based on the same technology as the RA53.

Cheers

Ian

[mediatechnology]

Ian -

Agreed. Those thermistors look to be in a class of the self-heating type rather than for temperature measurement which explains why they are in a glass insulating tube. They are also small to have low thermal mass. Unfortunately they may not be made anymore.

[mediatechnology]

Some small bead thermistors...

I think the key things to look for are the self-heating type suitable for liquid level or gas flow sensing.

https://www.ctscorp.com/Files/DataSheet ... asheet.pdf

Sensor Scientific small bead thermistors offer the advantages of small size and mass, very fast thermal response time, extreme sensitivity to current and voltage changes, and high stability.

https://atcsemitec.co.uk/wp-content/upl ... istors.pdf

https://f.hubspotusercontent40.net/hubf ... 14-web.pdf

https://www.mouser.com/datasheet/2/187/ ... 070802.pdf

Interesting paper:
https://www.qtisensing.com/wp-content/u ... istors.pdf

[ruffrecords]

Once again, thanks for the great references. So it seems a small bead type might be a possibility. Clearly these are aimed at temperature measurement so there is very little information about current sensitivity which is what we are after. I probably need to buy some and characterise them for this application.

Cheers

Ian

[mediatechnology]

Clearly these are aimed at temperature measurement so there is very little information about current sensitivity which is what we are after.

I think the metric to compare sensitivity is the power to raise the bead by 1°C in air.

For the RA53 it is the 'k" factor which is 12.5µW/°C.
It is a super-sensitive sensor.

Some manufacturers call it dissipation factor.

The other small beads I looked at are higher, such as the Amphenol/Thermometrics P20, are 140µW/°C.
The CTS are 100µW/°C with a fast 1s response time.

Perhaps one could scale the resistor/thermistor values and output level to get the current higher in order to use the higher dissipation factor parts.

Looks like the RA53 is run around 1 mA to obtain approximately 750Ω nominal.
The tau of the RA53 is 11 seconds.

[brianroth]

Being a Contrarian here....lol

What about a function generator? I've been using one for decades to align tape machines. Ruler flat output level from "DC to blue light".

Near instantaneous settling time when changing frequencies vs watching the many seconds of a HP oscillator's level bounce up and down when switched to the low frequency range.

"Ohhhh....but Brian! The THD is a percent". I'm not measuring THD....just aligning a tape machine.

Little known fact told to me by Jay McKnight (RIP). For decades, MRL has used a function generator to make the alignment tapes.

Bri

EDIT....the dedicated func gen chips from Exar and one other company are long gone. I think a (low power?) quad opamp could be used. Two sections (?) for the square/triangle and one for the diode bender to make the "sine" wave. Only power consumption issue would be if it has to drive +24 dBm into a 600 Ohm load.

[ruffrecords]

I understand exactly what you are saying Bri, but I am measuring distortion. I recently completed a 12 channel line level tube mixer. It is now doing its thing somewhere in Alabama. It was an interesting one, not just because it was line level only, but also because the customer wanted plenty of tube and transformer tone. Long story short, I needed to test the distortion at various levels and frequencies but I could not use my pocket oscillator because its distortion was too high; hence my interest in using a thermistor. I use it a lot because it is battery powered, small, light and has a calibrated 0dBu and -40dBu output levels. Its only problem is its distortion.

Spoiler alert: all I want to do is improve the one I have because I have plans for a much more versatile replacement

Cheers

Ian