Hello,
I got my hands on one of Pipelie's LNA's and made some measurements of some LTZ1000's.
I did a more in-depth write-up here: https://github.com/cellularmitosis/logs/blob/master/20180324-ltz1000-1f-noise/README.md
I had a question about loading the LTZ circuit. When discharged, the 1000uF input capacitor in the LNA appears as a 2k load to the LTZ, which would initially draw 3.5mA (decreasing to 1mA after about 3 seconds). Is it safe to draw this current from an LTZ? My intuition says yes, because a zener regulates a load by diverting current away from itself, so there is actually 3.5mA less current flowing through the LTZ during this period.
A variation on that question: if the LTZ nominally has about 4mA flowing through it, would there be any adverse effects of attempting to draw more than 4mA from the circuit?
Edit: initially my noise floor wasn't usable. using my aluminum dutch oven trick solved that.
I had a question about loading the LTZ circuit. When discharged, the 1000uF input capacitor in the LNA appears as a 2k load to the LTZ, which would initially draw 3.5mA (decreasing to 1mA after about 3 seconds). Is it safe to draw this current from an LTZ?
and it won't damage the lt1013 or LTZ,
Just a quick note.
With an LTZ1000A device and 15V on the collector of the heater transistor, the chip could heat up 300 deg C above ambient, worst case. If the heater transistor collector voltage is limited to 8V the worst case temperature rise is approximately 80 deg C above ambient. The actual temperature rise will depend on heater resistance, length of LTZ leads, PCB trace size, local airflow and several other things. A small Zener in series with the heater transistor collector can go a long way to prevent damage to the expensive chip.
The LTZ1000 chip does not get that hot because it has better heat conduction between the chip and the case.
I usually run a separate power supply for the heater transistor until I'm finished developing a circuit.
A small Zener in series with the heater transistor collector can go a long way to prevent damage to the expensive chip.
CM had huge troubles with that little zener
Take the blue pain killers when you want to touch the hot end of a 400W soldering iron,
R2 is a rather low value resistor. So the amplifier will have a low input impedance - with some sources this could lead to a reduced amplitude. This includes the case when the R1 for protection is active. In addition R2 would increasethe noise at around the lower frequency range.
From the noise perspective it is better to not have the input RC to set the lower frequency limit, but have this filter function at a later stage, after some amplification.
So except for faster settling there is not match advantage of a low value for R2.
For conditioning it might be a good idea to have a way to apply something like a 9 V battery to the input cap via a high value resistor. It might take quite some time under voltage to get the really lowest noise from electrolytic caps. This could be a real problem with bipolar caps though. So I am not sure they are a good idea.
The filter stage around U2 could be higher impedance in some points. So one could get away with lower capacitance for C12-C14. For a good filter a proper 2nd order active filter should give a better response curve than just a bunch of 1. st order RC combinations. The last filter stage directly at the output is also a problem, as the input impedance of the scope or what ever is connected would influence the bandwidth and gain.
The LT1012 (like other slow non AZ OPs) does not need that much decoupling - 100 nF + 100 µF would be already plenty. The extra decoupling caps make sense only if they are really close to the chip . so with the current layout the extra 33 nF and 100 nF caps at U1 do not help and may do more harm than good.
It might help to have access to the output of U1. This could be used for 2 purposes: check the input current / settling and for use as a wider bandwidth (e.g. 0.02 Hz - 10 kHz) amplifier, e.g. with a scope.
After reading your post...I decided to run some SPICE simulations of frequency response and changes in R1 and R2. Results are shown below. Based on this I will likely raise R2 to 100K and leave the jumper J3 out for most uses.
Note: I managed to find non-polarized electrolytic capacitors for the 1000uF and 2200uF coupling capacitors (Rubicon I think).