I need help from some one actually owning an LCR meter capable up to 1MHz.
Recently I was had to create some filters which them would fight capacitive load.
And I found that all diodes them are measured regarding their own capacitance at 1MHz.
My LCR this can do 1K up to 100KHz , and what I need to find out is capacitance VS test frequency of diodes this is in logarithmic or linear scale?
For example if one diode capacitance this is 120pF at 1 KHz, and becomes 50pF at 100KHz .. if by speculating this always become 4pF at 1MHz, then probably I can predict what to expect even by testing this diode at a lower frequency.
Recently I was had to create some filters which them would fight capacitive load.
And I found that all diodes them are measured regarding their own capacitance at 1MHz.
My LCR this can do 1K up to 100KHz , and what I need to find out is capacitance VS test frequency of diodes this is in logarithmic or linear scale?
For example if one diode capacitance this is 120pF at 1 KHz, and becomes 50pF at 100KHz .. if by speculating this always become 4pF at 1MHz, then probably I can predict what to expect even by testing this diode at a lower frequency.
Are we talking forward or reverse biased diodes?
I'm afraid what you're measuring there is not just the capacitance, but the Z of a more complex circuit. So depending on the frequency you would be giving more importance to one element or the other. Estimating the capacitance alone starting with these measurements is a rather convoluted affair.
I'm afraid what you're measuring there is not just the capacitance, but the Z of a more complex circuit. So depending on the frequency you would be giving more importance to one element or the other. Estimating the capacitance alone starting with these measurements is a rather convoluted affair.
I did measurements to component level.
The requirement of the circuitry this is me to use 1N914 that is 4pF at 1 MHz.
I have lots of other diodes, and I wanted by measuring them to find one in equivalence with this 1N914.
I started measuring other fast diodes which I do have here and it did amazes me their differences in capacitance.
The requirement of the circuitry this is me to use 1N914 that is 4pF at 1 MHz.
I have lots of other diodes, and I wanted by measuring them to find one in equivalence with this 1N914.
I started measuring other fast diodes which I do have here and it did amazes me their differences in capacitance.
All of the diodes that I've tested, have a reverse-bias junction capacitance which is independent of frequency. For these diodes, reverse-bias junction capacitance varies quite a bit with applied (reverse) voltage, and also varies with junction temperature, but not with frequency.
Be sure that you're measuring capacitance and not total impedance; you don't want the bulk resistance or the lead + fixture inductance, to pollute the measurements.
By the way, here's a good looking used HP LCR meter that goes up to 1 MHz, at an attractive price: link to eBay
Be sure that you're measuring capacitance and not total impedance; you don't want the bulk resistance or the lead + fixture inductance, to pollute the measurements.
By the way, here's a good looking used HP LCR meter that goes up to 1 MHz, at an attractive price: link to eBay
I am measuring capacitance, an LCR bridge this it does offer several measurements instantly, and I am not interested about impedance.
What I am trying to avoid, this is spending money for one LCR at 1 MHz.
What I am trying to avoid, this is spending money for one LCR at 1 MHz.
Diodes are measured at 1MHZ, because it is a convenient industry standard, but the capacitance does not primarily depend on frequency, unless you operate at extreme frequencies or with unusual components.
The cap-variation effect is almost mechanical, except that the input stimulus uses the same port as the output one, but other than that, they mostly behave like ordinary capacitors: at very high frequencies, the highest impedance parts will tend to vanish and reduce the useful capacitance (just like a pure passive cap).
For a 2:1 or even a 10:1 variation, this is immaterial.
You also have to measure diodes with a small signal voltage. Many capacitance meters use large signals (to generate a large current with small capacitors) which are unsuitable for diodes. Keep the AC voltage to 10mV rms. You might be able to build a 1MHz oscillator (LC type) with a low voltage output winding on the tank suitable for amplifying. You could feed the 10mV into the diode and measure the voltage across a sense resistor which only needs to be about 3.9k (one tenth of a 4pF capacitor) but then the voltage will need amplification to measure. (Expect 1mV). If you bias the diode with DC the capacitance will change of course.
This is where you are wrong, slow diodes them are made to serve at 50-60Hz , them is not our topic.
The all conversation is about FAST switching diodes, them they get tested at 1MHz.
And now lets talk with facts, diode 1N5819 this tested at several frequencies.
100Hz this measured 108.56 uF
120Hz ...... 118.73 uF
1K ..... 10.570 uF
10K ..... 1102 nF ( about 1uF + )
100KHz ... 105.55 nF ( 1/10 of 1uF)
Now is it possible someone to measure this specific diode at 1 MHz?
Additionally I will say that my LCR this using 0.5V when measuring components, this is ideal even for diodes because at the XY charts = capacitance VS reverse voltage, this specific 1N5819 diode capacity stays unaffected.
But from 0.7V & 0.8V this reverence voltage it does start influencing diodes capacitance.
But from 0.7V & 0.8V this reverence voltage it does start influencing diodes capacitance.
Quite. You are measuring forward bias capacitance if the voltage gets anywhere near 0.5V. This is a function of the charge in the diode (forward current x transit time), so you have to keep the AC signal small. You can alter the DC bias to obtain a C-V characteristic, but at high biases there is a high capacitance (especially in slow diodes where tt is long) and in addition the low impedance may affect the CV meter!!!
"120Hz ...... 118.73 uF"
If you really believe that is a real measurement, you need to go back to school!
A diode with 100microfarad capacitance??
If you really believe that is a real measurement, you need to go back to school!
A diode with 100microfarad capacitance??
Indeed.
I had made an honest attempt at answering, but Kiriakos being what he is (and has always been), it looks like a waste of time...
Either the leakage current of the schottky upsets the measurement, or he measures with a positive bias meaning he measures the diffusion capacitance, but that seems unlikely, because a non-specialized instrument is not going to provide a direct reading of diffusion capacitance (except perhaps by chance).
To make things clear: reverse capacitance for any type of diode will differ very little between 1KHz and 1MHz, provided the bias and measurement voltage are kept identical.
Any deviation from this points to a measurement error of some kind
Elvee
A capacitance meter using a high signal swing will measure some rather undefined combination of zero bias, forward bias and reverse bias capacitances. As forward is much larger I suspect that this is what is being measured. The meter will probably record the average of some forward bias value, which might be anything, and the very high values may indicate very low impedance (I would say short circuit but nothing is ever quite).
A capacitance meter using a high signal swing will measure some rather undefined combination of zero bias, forward bias and reverse bias capacitances. As forward is much larger I suspect that this is what is being measured. The meter will probably record the average of some forward bias value, which might be anything, and the very high values may indicate very low impedance (I would say short circuit but nothing is ever quite).
Yes the measurements are nonsense. All 1N914s and 1N4148 have less the 4 pF of capacitance as per their specifications.
Hard pressed to think of any small signal switching diode much above that. For a rough guide current handling should be inversely proportional to junction capacitance.
Hard pressed to think of any small signal switching diode much above that. For a rough guide current handling should be inversely proportional to junction capacitance.
Hi Ed, member Kiriakos says in post#10 that he is measuring an 1N5819 diode, not a 1N914 or 1N4148. I attach the capacitance-vs-voltage curves from several different manufacturer's 1N5819 datasheets, below. To find out which mfr's datasheet corresponds to which image below, hover your mouse over the thumbnail. The mfr name is in the image name.
These (Schottky) diodes are about 100-200pF at zero bias. And their transit times, being Schottky, are incredibly small.
For a rough guide current handling should be inversely proportional to junction capacitance.
I would have thought that physically bigger diodes had bigger capacitance AND bigger current handling. I.e. they are directly proportional, not inversely.
_
Attachments
Last edited:
My bad meant switching time/capacitance not current! Of course you need more junction size for more current. It is a Saturday will wake up eventually...
Last edited:
You need to go back to kindergarten because any measurement far and away from the specified by the manufacturer it self as product test frequency, all numbers them are invalid.
As more close to the truth I do consider all measurements from 1KHz and above.
- Status
- Not open for further replies.