500VA 240:35+35Vac
Vrso=28.4mVac
Vpso=241Vac
Vsso=74.3Vac
Rsso=1r0
N = 240/74 = 3.243
did the Calcs shown on the sheet by hand but used N=6.486 (for a 37Vac single secondary)
k = 0.999786
L11 = 0.0058
L12 = 0.00014
Lm = 26.9942
Are my calculations correct with that N value? Or do I need to use 3.243
What do the numbers mean?
What do they tell me about my transformer?
Are the numbers any good for these simulations with the 120VA transformers.
I don't have any small 35+35Vac transformers.
I do have a variety of small 25+25Vac, 225VA, 300VA and 30+30Vac, 120VA, 160VA,transformers.
Vrso=28.4mVac
Vpso=241Vac
Vsso=74.3Vac
Rsso=1r0
N = 240/74 = 3.243
did the Calcs shown on the sheet by hand but used N=6.486 (for a 37Vac single secondary)
k = 0.999786
L11 = 0.0058
L12 = 0.00014
Lm = 26.9942
Are my calculations correct with that N value? Or do I need to use 3.243
What do the numbers mean?
What do they tell me about my transformer?
Are the numbers any good for these simulations with the 120VA transformers.
I don't have any small 35+35Vac transformers.
I do have a variety of small 25+25Vac, 225VA, 300VA and 30+30Vac, 120VA, 160VA,transformers.
Thanks, Andrew! I will try those parameters in the simulator as soon as I can.
You did it correctly, for the single-secondary model.
I will not use your model parameters for a 120 VA transformer model. At first I will try it as is, i.e. as a 250VA secondary of a 500VA transformer, with the given voltages etc, and compare its behaviors to those of my derived scalable model.
Then I will try to per-unitize it, as I did for my original model with Terry's guidance, to make it scalable. It should be very helpful and enlightening to have a scalable model that was derived from a larger VA-rated transformer.
Tom
You did it correctly, for the single-secondary model.
I will not use your model parameters for a 120 VA transformer model. At first I will try it as is, i.e. as a 250VA secondary of a 500VA transformer, with the given voltages etc, and compare its behaviors to those of my derived scalable model.
Then I will try to per-unitize it, as I did for my original model with Terry's guidance, to make it scalable. It should be very helpful and enlightening to have a scalable model that was derived from a larger VA-rated transformer.
Tom
Last edited:
Daniel,
Paralleling reservoir caps and paralleling decoupling caps are not primarily about having less noise. In either case, it's mainly about enabling the current to flow more-easily and thus more-accurately, when called upon, because the impedance of the supply will be lower with paralleled caps. Since the current, in both cases, directly becomes the sound, paralleling more reservoir or decoupling caps (while not necessarily increasing the total capacitance) should always tend to give more-faithful reproduction of the sound, with lower distortion (steady-state or transient distortion or both).
If good layout practices are used, then the only reason, that I can see, to limit the number of paralleled caps, would be running out of space for them.
Unequal charge and discharge rates should not be a problem. But if they were, then increasing the number of paralleled caps seems like it should mitigate any associated effects, on average.
It is difficult for me to imagine how adding caps in parallel could create any noise, unless maybe the layout/current-routing is problematical, possibly with regard to star grounding. Or it might be possible that they are exposing a problem with some other part of the circuit.
Regards,
Tom
Last edited:
I'll just step in for Tom here, since he hasn't responded ...Hi Tom,
we all know that caps labled 4700/50volts are not exactly 4700ufd so that paralllelling them is actually paralleling several uneqaul value caps....did you simulate for those effects as well?
Somewhat unequal values makes absolutely no difference whatsoever, even if they are film caps. There is no such thing as matching, or anything like that coming into play here. If you happen to have a 2200, 3900, 4700, etc, on hand you can happily parallel them all up, especially if they're electrolytic, with absolutely no effect either one way or the other. The only thing that matters to any degree is the instrinsic parasitic characteristics, ESR and ESL, which can vary per the one capacitance rating. For example, Panasonic FCs can come in two different shaped cans for a certain value, and those two will have different ESR and ESL!
Frank
Hi G,
looks like my head is not in one place just now.
Yes, I have those numbers.
Vpss = 4.8Vac
Vrss = 0.22Vac
Rsss = 1/6 r (6 1r0 in parallel)And they run cold, you don't need a high power resistor, and you don't need to remeasure the resistance while it is still hot !
Shame it took 3 posts for the same test results.
looks like my head is not in one place just now.
Yes, I have those numbers.
Vpss = 4.8Vac
Vrss = 0.22Vac
Rsss = 1/6 r (6 1r0 in parallel)And they run cold, you don't need a high power resistor, and you don't need to remeasure the resistance while it is still hot !
Shame it took 3 posts for the same test results.
Last edited:
Sorry for being late,
so, aside from layout constraints, we would have no limit for the size of the "local" bank of caps. Do I understand correctly?
Stefano
Last edited:
No other UPPER limit for the NUMBER of paralleled caps, at least, as far as I can see, for either reservoir caps or decoupling caps.
There is usually a lower limit on the total capacitance, for PSU reservoir caps and for decoupling caps.
And there might be a lower limit on the number of paralleled caps, when using them for decoupling (i.e. near the point of load), after one has calculated the maximum tolerable inductance for the decoupling capacitance network.
Thanks for feedback,
Stefano
You are welcome, Stefano. In my last post I also wanted to clarify that by "size" we were probably talking about the number of capacitors being paralleled, and/or the physical size of their layout. "Size" could have also referred to the total value of the capacitance but I think that we were probably considering the case where we had already chosen a total capacitance and were then thinking about how many parallel caps to use to physically implement that value.
Stefano,
You might be able to find some difference, but, if you look at the calculations that I started to post, for decoupling caps, the examples were for high-power audio output devices, because it seemed like it was difficult to get sufficient decoupling with low-enough inductance, for many of those, to even be able to meet their specs.
I actually never (yet) took the time to do the calculations for something like an audio opamp circuit.
Your opamps probably already have something like the standard 10 uF || 0.1 uF from their power pins to ground. It would be interesting to run through the decoupling calculations and find out if the cap values used are sufficient, and if they are sufficiently close to the pins or not (according to the tolerable inductance you calculate).
Those decoupling cap calculations are all at some of the links I gave near the start of this thread, in post 27, at http://www.diyaudio.com/forums/power-supplies/216409-power-supply-resevoir-size-3.html#post3097232 . Maybe you could try it with "delta V" set to 5% of the rail voltage, at first.
Here are the three most-relevant links, from post 27:
http://www.diyaudio.com/forums/powe...lm-caps-electrolytic-caps-23.html#post2806854
http://www.diyaudio.com/forums/powe...lm-caps-electrolytic-caps-26.html#post2822959
http://www.diyaudio.com/forums/power-supplies/208579-30vdc-10a-psu.html#post2942537
Please NOTE that in most or all of those calculations, I used a wrong assumption for nH per inch, which was too low. The accepted standard guess is 25 nH per inch (or 1 nH per mm), for the self-inductance of "typical" wires or PCB traces. Or, you could calculate a better value for the actual conductor size and cross-sectional area.
Edit: The 0.1 uF (or other small value, probably X7R ceramic) that are at each opamp power pin are probably there for BOTH decoupling and bypassing. Bypassing is to bypass the high frequencies to ground, to defeat the positive feedback loop through the power rails for high frequencies that exists for most topologies of transistor-based amplifiers. Decoupling is for supplying fast transient currents when needed, so that the rail voltage is not disturbed too much, i.e. "decoupling" the effects of transient current draws from the rest of the power supply circuitry.
Cheers,
Tom
Last edited:
Thank you Tom,
<OT>
funny enough those opamps actually have no decoupling caps, at least in that model. The upper model had them (even if I can't understand why with such a high voltage rating (100v) with "reservoir" of 47/50v)...
</OT>
So, I'm just figuring out which values to use, trying to follow also Frank's rule of thumb on decoupling.
Ciao,
Stefano
<OT>
funny enough those opamps actually have no decoupling caps, at least in that model. The upper model had them (even if I can't understand why with such a high voltage rating (100v) with "reservoir" of 47/50v)...
</OT>
So, I'm just figuring out which values to use, trying to follow also Frank's rule of thumb on decoupling.
Ciao,
Stefano
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.