Hi Victor,
Scott is going to investigate this. So I think I will leave it him.
I tried this out of curiosity.
Cheers,
Vivnic is right when I first proposed this there was a cap, mainly for non-linear C. Since the DC gate current is negligible the R does not figure in. I have interest in this so I will mock it up and make a measurement.
A3U4 should be A4U4 and A3U2 should be A4U2 and A3U1 should be A4U1.
R1 refers to A4R1.
Last edited:
I have pare of those dual jfets Scott recommended.
I might give those a try.
They are very tiny don't know if my eye are good enough to solder them.
I might give those a try.
They are very tiny don't know if my eye are good enough to solder them.
Yes, the dual FET thing looked good but does not seem to help. Series/paralleling the FET's you have to get just the right Vp/Idss relationship looks better.
Last edited:
I see Glen Kleinschmidt has completed his audio THD analyzer with some remarkable resilts.
An Audio T.H.D. Analyser
An Audio T.H.D. Analyser
I think you need to set up a bridge circuit to measure them. With a little tweaking you could see the time constant from any two resistance points on the lamp pretty easily. The other secondary effect of the lamps would be the reactances at high frequencies. I don't know how inductive a filament would be. A quick check of a 6S6 lamp (12V 6W) I got 100 nH at 100 KHz. However different manufacturers will be very different.
Hi Demian,
I have an ESI impedance bridge 250DE which I haven't used yet.
I would have to dig out the manual for it. Perhaps I can use this.
I can definitely use it to get the inductance of the lamp but I'm not sure how I can use it to get the tau.
Cheers,
Hi David -- I would use a current source -- a biggish resistor -- in series with the bulb and look at the DC voltage run-up across the bulb with a scope after connecting the DC source voltage to the resistor. Choose a source voltage and final voltage on the bulb that puts the bulb near where it would be in the circuit. Pretty good way to check, I think.
BTW -- the 250DE tests internally at 1kHz, and is fairly useful with an external source up to around 5kHz -- but the resistance of the filament is going to give some strange readings when measuring the L. Bulb inductance will only matter up toward 100kHz, and some nHs won't cause serious effects, I don't think.
I was proposing a Whetsone bridge balanced to the target on resistance of the bulb and apply enough power to get the lamp up to temp and read the difference across the bridge nodes. It becomes much easier to see the change.
I used an ESI Videobridge to measure the inductance. I think it will be small and inconsequential.
I used an ESI Videobridge to measure the inductance. I think it will be small and inconsequential.
Hi Guys,
What's important here the time the lamp take to change and settle with a fixed frequency
sine changing from one level to another or a DC level with an AC component, there is a thermal inertia to over come. One can see this by observing the lamp intensity when an AC or DC step in amplitude is applied. The intensity is unchanged for a moment and then changes with the lamp's thermal tau.
When trying the multiplier with the 339a oscillator the AGC over steps then reverses and over steps again. It could be that the proportional gain in the the 339a AGC is simply too high for a lamp based multiplier because it's setup for a fast acting jfet.
This was expected.
It's not in my interest to use this for a Bridged T oscillator. I was just studying the effect.
I can find all this through trial and error but it might get me into the ball park faster by taking some measurements and doing some calculation.
What's important here the time the lamp take to change and settle with a fixed frequency
sine changing from one level to another or a DC level with an AC component, there is a thermal inertia to over come. One can see this by observing the lamp intensity when an AC or DC step in amplitude is applied. The intensity is unchanged for a moment and then changes with the lamp's thermal tau.
When trying the multiplier with the 339a oscillator the AGC over steps then reverses and over steps again. It could be that the proportional gain in the the 339a AGC is simply too high for a lamp based multiplier because it's setup for a fast acting jfet.
This was expected.
It's not in my interest to use this for a Bridged T oscillator. I was just studying the effect.
I can find all this through trial and error but it might get me into the ball park faster by taking some measurements and doing some calculation.
Last edited:
I've been giving a lot of thought to the idea of using another lamp's intensity to measure the amplitude of the oscillator. The lamp's intensity is measured by some sort of optical means.
The output of the detector is dc and can be directly applied to control circuitry.
There are a few reasons why this is attractive. Using a lamp as a sensor the behavior is matched by tau and moment of thermal inertia. It has the same response as the lamps in the multiplier. There is no fundamental doubling of the output as there is with a rectified or RMS detected system. The lamps tau offers a low pass function. It is simple.
I will order one or two of the solar cell dip IC's. If there is enough intensity from the lamp to drive them I think it will work. The data sheet shows a linear response to source intensity and they are fast. If that doesn't work then perhaps an LDR or other optical sensor.
I will still build and test my ISH as I have the parts for that and I want to try an MDAC multiplier.
Cheers,
The output of the detector is dc and can be directly applied to control circuitry.
There are a few reasons why this is attractive. Using a lamp as a sensor the behavior is matched by tau and moment of thermal inertia. It has the same response as the lamps in the multiplier. There is no fundamental doubling of the output as there is with a rectified or RMS detected system. The lamps tau offers a low pass function. It is simple.
I will order one or two of the solar cell dip IC's. If there is enough intensity from the lamp to drive them I think it will work. The data sheet shows a linear response to source intensity and they are fast. If that doesn't work then perhaps an LDR or other optical sensor.
I will still build and test my ISH as I have the parts for that and I want to try an MDAC multiplier.
Cheers,
You are better off with one big diode to maximize the short circuit current and use a conventional TIA. Everlight has some well under a dollar with a nice broad spectral response.
http://www.everlight.com/datasheets/PD333-3C-H0-L2_datasheet.pdf
Last edited:
Actually I have two ambient light sensors.
I'm not sure if this is quite the same but it might work too.
Connecting an ohm meter to it and shinning a LED flashlight it seems to go from a few mega ohm to a few hundred ohms.
Last edited: