I take it back. trr is as much as 50% of the discharge time, which, though it won't blow up, is really dreadful. If you switch to some series-parallel LED strings[1] the voltages will be lower, and you could use a jellybean 1n5819 schottky, to good effect.
[1] 4x4 (four strings of four) would be nice.--James
Ok, I tried adding a 1 ohm resistor in series with the power supply to monotor inductor current with poor results. The inductor measures around 25 mH with about 1 ohm resistance in the wire.
The mosfet is driven from a squarewave generator (8 volts) and there is a 330 ohm resistor in parallel with the inductor to discharge the inductor.
With a 4 volt power supply input, the waveform at the junction of inductor and 1 ohm resistor is a typical LR or inverted RC curve. Totally non-linear.
Time period is about 600uS per cycle and half that is idle time. During the mosfet on-time, the current waveform rises slowly for about 50uS and then increases exponentionally to about 1.5 Amps at about 300uS. At that point, current is near the maximum of 2 amps
But the waveform is not linear anywhere along the curve. It looks just like an inverted RC time curve The current starts out slow and then rises exponentially the same way the voltage rises across a capacitor rises with a series resistance.
And the time doesn't make any sense since the current reaches a maximum of 1.5 amps in 300uS, and working out the current for a 25mH inductor at 4 volts for 300uS is only 48mA. I = ET/L = (4*.0003)/ 0.025 = 48 mA
What's wrong with my inductor, or test setup?
-Bill
I have had horrible efficiency with 1N4000 series diodes rectifying upper audio frequencies. But I have had good results with a diode with part number HER105 (a Diodes Inc. part) - I got those from DigiKey.
Do you know the saturation current for that inductor?
When I make a switchmode LED driver, I like less inductance, higher frequency (25-100 KHz), and diodes that do well at high frequencies (like HER105 which does fairly well at 25 KHz). With a higher frequency, an inductor of decently small size can handle more current if its inductance is less.
Otherwise, I would check for measurement errors. See if the frequency is audible and sounds like what you would expect. One error that I have made before is forgetting that the times on that multiposition sweep rate switch of an oscilloscope is time per major division (usually around a centimeter), not time per trace. See if you are mixing up microseconds and milliseconds on the scope. One function generator that I purchased before had the frequency range switch pointing one range "off" what it was actually set to. Double check the scope input voltage range. I would also put that current measurement resistor on an ohmmeter. Also, I would have the connection from the current sense resistor to the scope being the usual shielded cable or at least a twisted pair, with the resistor leads as close as possible to shielded cable or twisted pair, or a proper scope probe. The scope ground should connect to the circuit at only one point, preferably an actual lead of the current sense resistor, preferably with that lead going to one of the supply rails.
- Don Klipstein ( snipped-for-privacy@misty.com)
AIUI your setup is:
. Vcc . -+- . | . .-. . | | Rsense . | | 1 ohm . '-' . | . +------> to 'scope . | . )|| . )|| L1 . )|| . | . | . | |--+ . | Q1 . | |
P.S. I have a whole reel of 100uH, 0.88 amp inductors, d.c.r. = 0.12 ohms, if that gives you any ideas. I'd be happy to send you some--just say the word. --James
Because it has a inherent capacitance - and hence the normally quoted resonant frequency for inductors
How do you know? If you get above 330mT or so you will be in trouble.
There are not many white leds with a forward voltage as low as this at
20mA.
With a 4V supply, assuming say 100mV Vce for a decent bipolar for such a job -(look at the Zetex ZTX e-lines for example) the transistor on time to get 700mA inductor current is 4.49mS
**If** your peak inductor current is 700mA with the transistor turned on then the peak current through the led circuit will be 700mA when the transistor turns off (providing it turns off quickly).Assuming a discharge voltage of 56V then the discharge time of the (25mH) coil is about 310uS from 700mA.
This is discontinuous operation.
A 1n400(x) diode is just about useless for such an application. It is all to do with the frequency of the turn on and turn off edges- not how many edges there are per second. Try a UF400(x) instead. Ensure that sure you turn the transistor off very quickly else you will just dissipate a significant portion of your 'output power' in this. With a Mosfet this will require a large gate current. A bipolar may be a better bet but pull the base to ground very hard to get a rapid turn off.
Stop random playing around with frquencies, duty cycles, etc and do your sums so you have some idea of what is happening
--Then play around--
Yes, I might try that. But I still have trouble measuring a homewound inductor. If I use the MOSFET / Generstor setup, I get a reasonably linear ramp current of 100mA in 125uS with a 4 volt supply, indicatating the inductor is around 5mH. But, if I use a sinewave from the generator with a 2.7K resistor in series and the inductor in parallel with a 1uF cap, the tank circuit resonates at around 1 KHz. indicating the inductor is many times greater (about 25mH).
Why the big difference?
-Bill
But how long does it stay a linear ramp? That's the point of monitoring the current. You said earlier the waveform was exponential viewed on a larger scale. When does the current ramp waveform angle skyward, and how much current is flowing then? That's the saturation point of your inductor.
Dunno. Something's wrong. Either your measurement, the cap isn't
1uF, bad cap (what kind is it?), the applied waveform isn't what we assume (e.g., contains d.c.), you're still using those horrible diodes, a cracked core, making a swinging inductor, non-linear magnetics...Measure the duration of the longest, honest linear ramp you can get across the inductor, then we'll talk.
Cheers, James
Partial Update: Bill's (the OP) inductor was saturating. The core was a giant ferrite ring, which appeared to be the kind used for common-mode EMI supression transformers. When he backed off from 100 turns to 10 turns, the thing 'worked.'
I've outfitted Bill with schottky rectifiers and a much smaller 100uH
0.88A inductor, and am hoping he'll grace us with a report...James Arthur
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.