And you would have to use a peak responding meter to get E and I and some other method to determine n. For the pulse type waveform described, knowing E (peak) and the average value of e(t), you could determine n. Or use a scope to do it all.
Not necessarily. How will you know the diode conduction voltage (which we're assuming is constant)? Suppose you don't have a scope, just meters. The voltage across the diode varies with time with the same waveshape as the current through the diode, so the conduction voltage can't be measured with an average responding or RMS responding meter; you'll have to use a peak responding meter which is somewhat uncommon. If we don't have a peak responding meter, then what? We can measure the RMS voltage across the diode and the RMS current through the diode and multiply them.
An RMS ammeter is appropriate if an RMS voltmeter is also used.
Phil presupposes that you know the diode conduction voltage, which is not constant when a pulse current is applied; some sort of device to measure the peak voltage across the diode would be needed. If you have a DVM which can measure RMS (AC+DC) voltage, it almost certainly will be able to measure RMS (AC+DC) current, and that's all you need. No need to make peak measurements.
RMS volts * RMS amps RMS watts in the general case. It is true here, so long as the diode forward drop is flat during the pulses.
Why not? Should be for reasonably short pulses, maybe not for seconds-wide things where thermals start to matter.
Most everybody has an average-reading DC ammeter and a scope with a regular voltage probe. What's less common is a good, calibrated current probe for the scope, or a DC ammeter meter that does true RMS. So the suggested combo works nicely for the original problem, to figure the diode dissipation for rectangular pulses.
If diode drop varies significantly during the pulse on time, neither meter-only technique is accurate.
That's the assumption Phil made, and I made it too.
It is also important that the diode voltage drop to zero when it isn't conducting a current. If the diode were replaced by a battery so that the voltage was constant all the time, the RMS volts * RMS current wouldn't work. But, *in this particular case*, it does work.
Because by pulse current, we mean a current that is on for a while and off for a while. That's what I mean by "not constant", not that that the voltage while the diode is conducting isn't constant *during that time*. Sorry if there was confusion.
If you use a scope to measure the diode conduction voltage, then you only get scope accuracy, not the accuracy you could get with a DVM.
So it does. Everybody has a DVM that can measure both AC (RMS with or without DC) and DC current these days, I think, and the true RMS of AC+DC can always be done by hand with the calculator that everybody also has. :-) Just measure the voltage and current in both AC and DC modes and take the square root of the sum of the squares of the two voltage measurements and of the two current measurements. Then you get the result with only one meter.
Quite right. Then we need to use the wattmeter everbody should have, or the really neat new 4-channel scope that does trace math that everbody wishes they had. :-) :-)!!
Before I respond to specific items, let me say that all my comments have assumed that we were dealing with line frequency currents. The OP didn't say explicitly that he was concerned with line frequency rectifiers, but he *was* concerned with power dissipation in a diode. Typically, signal rectifiers aren't dissipation limited. It's usually line frequency rectifiers or, nowadays, diodes in switching power supplies where power dissipation is a concern. To keep things simple, I assumed line frequency. You may not think it was a reasonable assumption; if so, then we will have to differ.
Sure it can. Have a look at the URL John Larkin posted:
formatting link
All of the DMMs there have a MIN-MAX function which can easily get the diode on voltage at line frequency. As I said in a posting, you need a peak responding meter, and that's what MIN-MAX does.
I can't comment on your experience, but my own differs. Every DMM I have used on the job has had substantially greater bandwidth than a few KHz. I personally own three DMMs. The two handhelds, an HP 974A and a Fluke 189 both have 100 KHz bandwith. My bench meter is a 20 year old HP 3468 with 300 KHz bandwidth. The lowest bandwidth shown on the Fluke site referenced above is 20 KHz.
But this isn't relevant to my postings since I was assuming line frequency.
John Larkin first posited "...a clean rectangular pulse..." in this thread, something that will scarcely ever be seen in practice. In line frequency applications, one usually sees currents looking like pieces of sinusoids, or rounded pulses as in a capacitor input filter. In switchers, the pulse currents in diodes range from triangle waves to trapezoids, but are rarely flat topped. So neither method will work well in practice to determine diode dissipation. As I mentioned in another post, one would have to use a wattmeter (good to low audio frequencies, if it's an old fashioned dynamometer type), or a scope with trace math, or some other purpose-built piece of equipment.
But John's hypothetical is a simple case which you used to point out that the power delivered to a constant voltage by a varying current is proportional to the *average* value of the current. This is quite true, and, I dare say, often overlooked.
I merely pointed out that since the voltage across the diode isn't absolutely pure DC, but is a pulsed waveform like the current, the measurement can be made by measuring the true RMS voltage and current and multiplying the two.
This is an idealized situation and my comments were intended to be taken as applying to an ideal where accurate RMS measurements could be made, much as your method only works for the ideal situation where the pulse is flat topped.
John's description led me to think that his "clean rectangular pulse" might be coming from a pulse generator, and so I assumed that there was no reverse voltage across the diode. If there is, then it can be blocked by another diode; and if that diode's forward drop causes too much error, it can probably be compensated.
*Of course* it's true that if the measurements are in error because the waveform is at a frequency beyond the capability of the DMM, then this method won't work. I never claimed otherwise.
Hi, Fred! See another post I made in response to Phil. I did assume that there was no reverse voltage across the diode. Since John Larkin talked about a "clean rectangular pulse", I considered the simplest case where the driving signal came from a pulse generator and was only positive or zero. I assumed this, because in practice in most power handling circuits, you won't see a "clean rectangular pulse". Hence, it seemed to me that it was an idealized situation under discussion.
So, if the flat-topped current through the diode is I amps, and the "on" voltage is E volts, you can replace the diode with a resistor of value E/I ohms, and you will see exactly the same waveforms. And the dissipation in the resistor will be RMS volts * RMS amps, and also in the diode. This is only true if the voltage across the diode doesn't vary during the pulse.
If the voltage across the diode during the "on" time of the pulse doesn't change (remember the current is constant during the "on" time) due to heating or some other cause, then you wouldn't be able to tell that it wasn't a resistor (assuming no reverse voltage, as I said above). Although I suppose that if the voltage did change you could attribute it to heating of the supposed resistor. But it might change in a more "diode like" manner if it were really a diode; ah, well, one complication after another.
The op specified a "pulsed waveform" and wanted to calculate diode power dissipation. For either a regulator efficiency calculation or for sizing a heatsink, 20% accuracy should be good enough, so reasonable approximations should be OK. For the purposes of flame waring, there is no such thing as "reasonable approximations."
Only true in the case of a resistor. Make that a diode rectifying 0.5A average on a 1000V circuit and you meter wi'll make you need a pretty huge heatsink for you DO41 diode.
They're only bad if you want to measure RMS. As far as bandwidth goes, cheap averagers can be just as bad as cheap RMS circuits. Most do handle 60 Hz pretty well.
Few meters will measure DC-coupled RMS, on the voltage or, pertinent to this thread, current ranges.
I have a box full of flea-market thermal converters, which sort of look like acorn tubes with flying leads. Inside is a heater and a thermocouple. They were, maybe still are, used as ac-dc transfer standards.
Fluke and somebody else still make thermal-transfer based bench DVMs, very accurate and very expensive.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.