How low? Custom transformers can get pretty weird in form factors. Just remember the 1 Henry coil used in a dial-up modem for wet telco connections, audio coupling transformer that is about as flat as a thick credit card! From memory, 54 Awg wire, no coil form, run multi-amps through to heat the wire so the insulation eases and bonds into its own coil form. [temperature cycling probably kills this thing fast, though.]
spice files are text, simply scoop and post we'll sort out the wordwraps.
Perhaps not. We might be talking about thermosetting (SV) self-bonding wire here - some kind of two-part resin bonding agent where the two components don't react very rapidly at room temperature, but do react quite fast, and irreversibly, when you put enough amps through the coil to get it good and warm.
in your original post said 50kHz, not 15kHz AND 33uF, not 33nF Glad cause those caps sounded large. I'm assuming all typos and the spice net list is accurate.
not sure how to convert the V1 line, usually V1 n001 n002 description, so does that 5 mean 5Vdc? or AC=5, can't because doesn't appear in the SIN description. The .ac analyses usually don't work well for these types of circuits anyway. looks like I'll throw away that 5.
That netlist pretty much looks like what I did, except I used 33uF and
50kHz, but will work at it a bit.
meanwhile change that .tran 100mS to .tran 0 100ms 0 1uS and run it again. I bet you'll find the 'rumbling' goes away. If it doesn't tighten your options for closer/better solution before quitting and calculating the next point.
I set the final time as 1 Second so the minimum frequency would be 1Hz. It's difficult to see 2Hz with only 100mS run time.
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: skypeanalog | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
I posted a correction. The "rumble" is similar be it 15K or 50KHz. I was just playing with the values to see if it would go away. it doesn't. I did the 1uS thing and it did get substantially reduced. like you said maybe a simulation thing.
I'm just wondering of it is the diode biasing thing is why I'm seeing similar results with the lab setup. these have high Vf and if not all are biased on at the same time thats 20-30V of Vf that can modulate the output voltage. As an experiment I may try to load each tap to see if the output stabilizes.
Also, I found the problem with the apparent instability! at least think I have.
I let LTspice run on its own [only specifying STOP time] and had some wobbling, but less than 0.1Vrms, not that 2V reported.
THEN I forced the steps smaller, down to 5uS max [which as you and I both know is the only way to sometimes smooth down a .tran analysis] and then I actually got more than 2V wobbly!! say what?! upon inspection of the plot, found that the plot program itself decimates the data, making it really 'chunky' looking, ramps between data points were NOT 5uS, but rather anywhere from 50uS to 65uS and all over the place! Ran the EXACT same analysis again, but this time used .options plotwinsize=0 to prevent the plot program from chewing up the data and VOILA! the wobble dropped below 1uV !!! in the FFT, below -90dBV
So there are two-fold places to look when you see unstable answers. First place is to ask, "Are my maximum step sizes small enough?" and the second place is, "Does the plot program chew up my very smooth data?"
In this case, it was the plot routine chewing up the data and therefore injecting noise into the results!
I'm tempted to apply .tranoise just to see what happens in a more accurate siimulation.
But that will take over 42 independent noise sources [2 each for diodes and one each for source and load resistors], but would be fun to see the results.
you must have missed my reply to Jim Thompson posting the model for the CSD diode
=-=-= Also, I found the problem with the apparent instability! at least think I have.
I let LTspice run on its own [only specifying STOP time] and had some wobbling, but less than 0.1Vrms, not that 2V reported.
THEN I forced the steps smaller, down to 5uS max [which as you and I both know is the only way to sometimes smooth down a .tran analysis] and then I actually got more than 2V wobbly!! say what?! upon inspection of the plot, found that the plot program itself decimates the data, making it really 'chunky' looking, ramps between data points were NOT 5uS, but rather anywhere from 50uS to 65uS and all over the place! Ran the EXACT same analysis again, but this time used .options plotwinsize=0 to prevent the plot program from chewing up the data and VOILA! the wobble dropped below 1uV !!! in the FFT, below -90dBV
So there are two-fold places to look when you see unstable answers. First place is to ask, "Are my maximum step sizes small enough?" and the second place is, "Does the plot program chew up my very smooth data?"
In this case, it was the plot routine chewing up the data and therefore injecting noise into the results!
I'm tempted to apply .tranoise just to see what happens in a more accurate siimulation.
But that will take over 42 independent noise sources [2 each for diodes and one each for source and load resistors], but would be fun to see the results.
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.