Noise in "Jelly Bean" regulators - Part 4

I again revisited the concept of a low noise audio amplifier. The excellent Op Amp noise calculator suggested that the NF of my amplifier was 26dB. Now that got me thinking: pretty any random mmic has a NF of below 6dB. So why not use a mmic rated to work from DC upwards as a front end?

I didn't have much luck finding any prior work in using a mmic as an audio amplifier. But it did trigger the recollection that Direct Conversion receivers use low noise audio amps and that "Solid State Design for the radio Amateur" had used an all transistor amplifier in a conceptual DC receiver. 

That evening I sat in front of the TV with a calculator and during the commercial breaks reworked the biasing for use with a 9V battery and reduced the overall voltage gain to 1000. That should allow me to see input noise levels down to perhaps 5uV. I ended up with the following:

 

Unfortunately my amplifier was a very good oscillator when the input was shorted or terminated in a low ohm resistor. I haven't been able to stop the oscillations so a different amplifier is going to be built.

73's

Relay Life for Relay #2 - Degradation Found

While the abused relay performed surprisingly well (see here) it did start to degrade. After 6,449,686 cycles the contacts did not open for some 63 cycles. The evidence for this was the declining voltage across the contacts as the constant current regulator warmed up and current limiting set in.

After this point there were numerous "runs" where the voltage across the contacts was much higher than average. 

If relay performance is considered as good then degrading to failure, at what point has it failed? Apart from the periods of degraded performance the relay is still switching reliably at 8.4 million cycles.

In light of the degradation seen at 6.4 million cycles I went back through the data. If the symptom of contacts not opening correctly is a decline in the voltage across the contacts over time, then the relay's first, albeit brief, period of degradation started at 44,296 cycles. 

It would appear that the relay is still fit for the intended purpose of TX/RX changeover. Now it's time to sort our the isolation and loss characteristics.

73's