EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: ELS122 on September 12, 2023, 12:25:37 am
-
Like a tiny SMD diode and resistor combo, what's the settling time I could expect from that?
-
Depends on so many factors. Thermal mass of components, how much power is being dissipated, how much the power level has changed, etc. Are you making your own?
Here's a chart from a Keysight N8480 Thermocouple RF Power Sensor:
-
Hi,
Linear Technology´s LT1088 specs Bandwidth limits of 50MHz (for 1% accuracy) and 100MHz (for 2%).
The lower bandwidth Limit is defined by the servo loop of the circuit ... which LT defined to be 50Hz (below that freq the circit stated to follow the input waveform)
regards
Calvin
-
LT1088 settling time is on page 3 bottom. Around 400 ms depending how accurate you would like to be. First crossing is around 50 ms.
-
For the settling speed the distance between the heater and sensor really matter. Twice the size gives 4 times the settling time.
The time can well be in the 10s of seconds and may need several times constants to really settle.
-
For the settling speed the distance between the heater and sensor really matter. Twice the size gives 4 times the settling time.
The time can well be in the 10s of seconds and may need several times constants to really settle.
Well I plan to attach the diode and resistor as close as possible, as close as the dies are in 402 size packages can be at least...
I'd be happy with a 0.5s settling time to get within 5%, does that seem doable with diy diode+resistor combos?
-
Likely not going to be able to get 0.5 seconds response with that approach.
If you create a thermal model for the resistor and diode, the simplest is a shunt C for the 0402 for thermal capacity, then series R for thermal impedance of 0402, then series R, shunt C, series R for interface thermal impedance shunt capacity and impedance, then series R for diode 0402 and finally a shunt C for diode 0402.
This simple model shows that the propagating sense voltage is impeded by a series of simple RC low pass filters, each with significant delays, and doesn't even include the thermal capacity of the surrounding 0402 mountings, so expecting a 0.5 second step response settling time is wishful thinking!!
Eons ago Fluke utilized a clever RMS detector based upon pair of tiny transistors spaced apart (high thermal impedance) on a single die, each with a precision low TC thin film resistor placed directly over the transistor BE junction (low thermal impedance). This die was thinned to reduce thermal mass (capacitance) and die suspended within a hermetic case (believe TO-99) by the contact wire bonds to insulate to the outside ambient and reduce thermal mass (capacitance) effects from the outside.
This operated by the input signal heating the first thin film resistor (Vrms^2/R) which caused a change in the first transistor Vbe {Vbe = kT/q * ln(Ic/Is)}, this Vbe was sensed by a precision Op-Amp in a closed loop integrator fashion which drove the second resistor with a voltage Vout to produce the same Vbe on the second transistor. Thus Vout was equal to the input voltage heating effect, or it's RMS value!!!
Clever indeed, however all this effort still imposed a settling time in the fractions of a second!!!
Best,
-
There is also a difference in how the converter is used. Just waiting for the settling of the temperature is slow, especially when waiting for settling very well (e.g. 0.1% or better) as this includes the insulation / connecting wires to get slightly warm. The settling gets faster if there is a regulation loop to keep the temperature constant, even if this needs an additional heater or a superimposed signal to the same resistor. The advanatage here is the transisents are more confined to the center part and not such much effected by part further away.
-
I did that (before I got the AD8307 for measuring my VFO's and qrp TRXes power) with smd parts on thin wires (50ohm+diode), glued together with second glue, I put that into a small foam ball.
Basically it "shows something" but it depends on many factors related to the thermal dynamics of entire system, nobody can tell you, you have to simply try and report.. It will require some calibration, I did it with DC current over the resistor.. Long time back..
PS: it is done usually in a differential mode, where a second diode (of the same type/params) thermally apart of the rms sensing diode senses the ambient temperature..