Simple power, wasn't so simple.
Hey smart guys, I need your estimates for these additions:
Cap and resistor values for RC (or RCR) across each DC rail, and located near the bridge rectifiers.
Cap and resistor values for RC (or RCR) across each transformer secondary winding. X2 caps can be used if desired.
Data:
1). With this capacitance, I believe that one may assume 24.5+24.5vac dual secondaries transformer, probably toroid and an average 270va (range 240va to 300va), but the exact model of transformer used by the constructor cannot be predicted, so please aim for "mild" RC values that will in all cases be better than omission.
2). Given the habits of diy audiophiles the exact model of bridge rectifier diodes used by the constructor cannot be predicted (could be terrible MUR, could be fantastic Fairchild Stealth, might be KBPC1004), so due to vast variance, there is no snubbing of individual diodes.
3). The remainder of the available data is shown in the attachment below. Since the last time you saw it, some simplistic RF filtering has been added, capacitance has been upgraded to support a "normal" range of transformers and the resistors have been upgraded to non-inductive types. Now it needs well estimated RC's added before-and-after the bridge rectifiers, and I couldn't calculate the values for them. Help?
Hey smart guys, I need your estimates for these additions:
Cap and resistor values for RC (or RCR) across each DC rail, and located near the bridge rectifiers.
Cap and resistor values for RC (or RCR) across each transformer secondary winding. X2 caps can be used if desired.
Data:
1). With this capacitance, I believe that one may assume 24.5+24.5vac dual secondaries transformer, probably toroid and an average 270va (range 240va to 300va), but the exact model of transformer used by the constructor cannot be predicted, so please aim for "mild" RC values that will in all cases be better than omission.
2). Given the habits of diy audiophiles the exact model of bridge rectifier diodes used by the constructor cannot be predicted (could be terrible MUR, could be fantastic Fairchild Stealth, might be KBPC1004), so due to vast variance, there is no snubbing of individual diodes.
3). The remainder of the available data is shown in the attachment below. Since the last time you saw it, some simplistic RF filtering has been added, capacitance has been upgraded to support a "normal" range of transformers and the resistors have been upgraded to non-inductive types. Now it needs well estimated RC's added before-and-after the bridge rectifiers, and I couldn't calculate the values for them. Help?
Attachments
Last edited:
I typically use a total series R of 1x..3x the transformer winding DC resistance. With your scheme of diode//resistor I'd put (part of) that resistance in front of it to have some lossy resistance in series to the diode's high capacitance, and increase the parallel R to about 1R or even more (and perhaps two series diodes), depending on which specific soft-sag characterisic looks best for the application.
EDIT : Snubbers accross primary and/or diode bridge are probably best found empirically and by measurement. One may not need them, though, when series resistances provide enough losses to act simultaneously as snubbers.
EDIT : Snubbers accross primary and/or diode bridge are probably best found empirically and by measurement. One may not need them, though, when series resistances provide enough losses to act simultaneously as snubbers.
Last edited:
Daniel,
If you are talking about snubber networks, to damp unwanted resonances that might be excited by the diode turn-off, there is NO good generic recipe, because the resistance that does the damping needs to be set equal to the characteristic impedance of the specific resonant LC tank that needs to be damped i.e. Z = sqrt(L/C). And the L is probably/usually/often from the parasitic inductance in the transformer secondary, while the C might be from what the diode junctions become as they turn off.
The resistance does the actual snubbing. The C is there so that only the unwanted frequencies cause power to dissipate in the R.
Both values need to be determined AFTER the circuit is built, using a procedure similar to: http://www.diyaudio.com/forums/powe...lm-caps-electrolytic-caps-30.html#post2828689 .
Tom
Question(s):KSTR said:
Would you like to raise the resistor value up to 0.33R, and are 9w 0.33R resistors strong enough?
This is a beginners power supply, with led's that indicate to stay away from the lethal charge. Even when not aiming the build complexity higher than the moon, it seems possible to improve filtering because this is specifically a stereo power supply for specifically a pair of TDA7294 amplifiers, with specifically transformers from 240va to 300 va range. Although optimal might be nice, I'd settle for practical.gootee said:
Problem:
I don't have low-and-high examples from which to create a reasonable (albeit slightly imprecise) average for snubbing the transformer secondary windings.
Last edited:
P=I²R, so a 10W 0R33 is OK for 5 Ampere continuous current. Select a wirewound that has high peak/surge current capability (which is generally the case for most any wirewound) and mount them such as to NOT heat the PCB, that is, with clearance.
For example, I've used 4pcs 5W 0R47's around the diode bridge in a 2x50W active speaker (100W xformer) which gave enough wattage headroom for normal application, playing music signal at home at moderate levels. For a heavy duty laboratory amp which must deliver full spec'd power for any amount of time and take any overload with grace, one should beef up transformer and resistor wattages considerably (as well as increasing supply capacitance and heatsinking of the chips), like 3x...4x.
CRC resistors
When I checked the non-inductive resistor prices at Mouser, I was pleased by 4 parallel 1.3 ohm 3W metal oxide when teamed up with a 6a05 for surge current. Will these handle the current of stereo TDA7294? Or do the crc series elements need strengthened?
Question: Bass?
In order to compensate for the decreased bass impact consequences of changing the "crc resistor" value from 0.25R up to 0.33R, I'm quite tempted to use a fast silicon diode in order to get the switch-on point a tiny bit earlier for the "rail stiffener diodes," and the fast silicon is somewhat more quiet. When I try shopping, that results in FFPF06U20STU, FR603-B, and STTH802D types. Are these more suitable than the 6a05?
Rectifier noise
Also yet more research to dump the need of dc side snubbers by swapping to Stealth schottky diodes, but was concerned about the hazards of complexity, and therefore, some sort of prefab is needed. Then I discovered that the dual series package is conveniently same width as the rails, and the "confusion factor" is reduced considerably (there's only 1 package per each transformer output wire). Wow, nice low parts count. Higher surge tolerance and less voltage drop than KBPC1004's.
A new question:
Are 4 of these (datasheet) big enough for the job of stereo TDA7294?
Transformer inductance peak
(In some ways similar to woofer inductance peak)
If a prefab crossover were made for a limited range of similar woofers, the nonspecific filter would remove at least half of the peak, which is a lot better than allowing all of the peak. Much to the point, the end user would be "in the ballpark" prior to attempted fine tuning (if any).
Likewise. . .
Could you measure a 24+24 (or 25+25) 300va transformer for the values of the secondary windings snubbers? I simply need a "better than nothing but not worse than nothing" estimate. Like a rubber bullet, to get a non-specific snubber to work needs somewhat less resistor current in order to accomplish the wider tolerances. The location of the missing filters are at the spots marked RC (in the attachment below).
Question:
In the near-absence of available data, I did find a chart for this application, and the capacitor value in common to most of the likely transformers is 4.7n, and so if this were used, what's a reasonable (but slightly less than optimal current) resistor value to put with it for the RC to snub a transformer secondary?
Attachments
Last edited:
Transformer secondary's leakage inductance will depend on how it was manufactured. But ONE particular 25VRMS/300VA transformer I can get a number for has 43.3 uH there.
The snubber resistor value should be set to
Z = √(L/C)
I was thinking that the parasitic capacitance was dependent on the diodes. And it will also depend on the layout, etc.
But IF you were to use 43.3 uH and 4.7 nF, then the optimal snubber resistance would be sqrt(43.3u/4.7n) = 96 Ohms .
For the series capacitance in the snubber network, you should use 4 to 10 times the parasitic capacitance. Using 10x would give a little better damping by the resistor at the expense of more power dissipated in the resistor.
The series capacitor's value would be in the range of 18.8 nF to 47 nF, for your case.
My secondary has a 14KHz resonance. I used a 1.5u/22R snubber for this. I found there was a surprising increase in image depth after adding the snubber. I could not find any ultrasonic resonances which I falsely remembered but there was a resonance at 14MHz that was totally unaffected by the snubber.
This does not agree well with your analysis or mine. I discovered that mains impedance affected resonance greatly (I have a mains filter+switch and a variac), but the region several Khz and up was mostly unaffected and tunable via RC. I think trafo resonance may be highly dependent on external impedances, so it may not be practical to employ a secondary RC if a mains filter is not also used.
This does not agree well with your analysis or mine. I discovered that mains impedance affected resonance greatly (I have a mains filter+switch and a variac), but the region several Khz and up was mostly unaffected and tunable via RC. I think trafo resonance may be highly dependent on external impedances, so it may not be practical to employ a secondary RC if a mains filter is not also used.
For the sake of argument (did I just say that ? 
) the power supply redjr used in his build hasn't received any comment. As there have been two approaches discussed, the basic kit upgrade and/or the new advanced PCB/PS, does anyone have an idea/recommendation or two on that unit they can share?
SMPS300R
) the power supply redjr used in his build hasn't received any comment. As there have been two approaches discussed, the basic kit upgrade and/or the new advanced PCB/PS, does anyone have an idea/recommendation or two on that unit they can share?SMPS300R
I still wonder if a standard bridge rectifier, like the KBPC1004 has a predictable reaction common to the likely range of transformers and therefore could be snubbed approximately? For example, it is popular to add 4.7n X2 across the bridge (across the secondary), often, but not always, resulting in an audible improvement. Although the cap might make the transformer perform slightly worse, it has done the bigger job of confining the bridge rectifier noise away from the transformer. So, my question is how to get this particular job done somewhat more reliably with an RC?
Maybe it is possible to use RC values unlikely to make the transformer perform worse, but yet still manage to confine/muffle bridge rectifier noise to keep some of that out of the transformer? In this way, an RC seems more reliable/predictable than a "just the cap" approach.
In my opinion, the KBPC1004 prefab bridge rectifier is ideal for beginners projects because it is clearly marked and non-confusing to install.
Question:
Is there a recommendable, secondary snubbing, estimated RC that can make the KBPC1004 perform as nicely or better than a botique (muffled) bridge rectifier?
Maybe it is possible to use RC values unlikely to make the transformer perform worse, but yet still manage to confine/muffle bridge rectifier noise to keep some of that out of the transformer? In this way, an RC seems more reliable/predictable than a "just the cap" approach.
In my opinion, the KBPC1004 prefab bridge rectifier is ideal for beginners projects because it is clearly marked and non-confusing to install.
Question:
Is there a recommendable, secondary snubbing, estimated RC that can make the KBPC1004 perform as nicely or better than a botique (muffled) bridge rectifier?
Last edited:
I have a stupid idea, and no means to measure it. But, if there's merit, it probably needs checked out.
A series pair of 10n ordinary polyester capacitors is a 5n with double the loss. This is a cap at high pitches, and effectively an RC at low pitches.
If this were located upon the KBPC1004 from "~" to "~" then would it:
"muffle the power, not the audio?"
Well, that's the point.
A series pair of 10n ordinary polyester capacitors is a 5n with double the loss. This is a cap at high pitches, and effectively an RC at low pitches.
If this were located upon the KBPC1004 from "~" to "~" then would it:
"muffle the power, not the audio?"
Well, that's the point.
Last edited:
MKT is still relatively lossless. Now a 3.3u NP lytic? That's an idea. Assuming a NP lytic could handle the AC voltage without reduced lifespan, a relatively large 10u or so with series resistor could be highly effective. My 1.5u+16R snubber is only almost large enough to damp the 14KHz resonance. If I put R too low I end up with a 5KHz resonance.
You guys do realize that the _resistor_ is the snubber, right? And you realize that the optional series capacitor is only used to try to ensure that only the unwanted frequencies get exposed to the snubbing/damping/terminating resistor, in cases where there would be too much power dissipated by the resistor otherwise, right?
You COULD "possibly" use an electrolytic, or other capacitor, by itself, IF you happened to find one that had an ESR that was equal to the characteristic impedance of the LC resonance that was causing the ringing or oscillation problem. But that would be extremely rare, I imagine. In every other case, a capacitor by itself is NOT a snubber.
Using a large electrolytic in any other case will likely just make the snubbing resistor get hotter than it needs to get. Or are you trying to make something other than a snubber?
You COULD "possibly" use an electrolytic, or other capacitor, by itself, IF you happened to find one that had an ESR that was equal to the characteristic impedance of the LC resonance that was causing the ringing or oscillation problem. But that would be extremely rare, I imagine. In every other case, a capacitor by itself is NOT a snubber.
Using a large electrolytic in any other case will likely just make the snubbing resistor get hotter than it needs to get. Or are you trying to make something other than a snubber?
Gootee, my point was that my resonance was at 14KHz, and a 1.5u film is still too small to effectively snub it without causing another 5KHz resonance. I mistakenly construed that the cap was the problem. And yes, a lytic could of course be used with the appropriate snubbing resistor, it's just I've been around DIYAudio too long and am lytic phobic . And in this case we don't really need to worry much about the ESR as long as its reasonably low, we can just use a series resistor like we do with film caps.
. And in this case we don't really need to worry much about the ESR as long as its reasonably low, we can just use a series resistor like we do with film caps.
Still looking forward to the possibility of an answer to this:
Is there a recommendable, secondary snubbing, estimated RC that can make the KBPC1004 perform as nicely or better than a botique (muffled) bridge rectifier?
(trying to make a smaller request there)
In comparison, MUR860 can muffle both the AC side and the DC side, while also being very unkind to radio reception and possibly irritate the audio amplifier or at least make extra work. These and other random consequences indicate that "botique bridge rectifier" is not paradise. So, I'd much rather muffle just the AC side of a KBPC1004--That looks a bit more sensible and maybe less likely to randomly nuke a tuner. People who live at fringe range reception areas wouldn't appreciate having to turn off the power amp before listening to the radio. So, it seems that the main task is to confine the bridge rectifier's racket to the locale of the bridge rectifier.
Is there a recommendable, secondary snubbing, estimated RC that can make the KBPC1004 perform as nicely or better than a botique (muffled) bridge rectifier?
(trying to make a smaller request there)
In comparison, MUR860 can muffle both the AC side and the DC side, while also being very unkind to radio reception and possibly irritate the audio amplifier or at least make extra work. These and other random consequences indicate that "botique bridge rectifier" is not paradise. So, I'd much rather muffle just the AC side of a KBPC1004--That looks a bit more sensible and maybe less likely to randomly nuke a tuner. People who live at fringe range reception areas wouldn't appreciate having to turn off the power amp before listening to the radio. So, it seems that the main task is to confine the bridge rectifier's racket to the locale of the bridge rectifier.
Perhaps 15uF+10R would do it. This sets the corner at 1KHz, low enough probably to get both my 14KHz and 5KHz resonances (hopefully not creating a 1KHz resonance). If your secondary output was over 28VRMS you would need more than a 1/4W resistor. I believe 28VRMS~39.5VDC? In any case I would choose a low-ESR lytic if possible. If the ESR of the cap shares a significant voltage with the R, then the cap will be dissipating a share of the heat, and we don't want lytics self-heating.
So, what could you do with a 2u2 polyester like http://www.mouser.com/ds/2/315/ABD0000CE29-44537.pdf And, they are rated safe for this application.
Technics production units had 10n caps per each diode (pair of specialty 3-pin ceramic caps internally like "pin-10n-pin-10n-pin"), and Mark Houston's gainclone has 10n polyester dip caps per each diode. Both cases, measured unloaded, produce about 1v to 2v less junk after snubbing the diodes. That looks silly expressed in volts; however, it looks impressive if expressed as a percentage. Smaller scale amplifiers are sometimes shown with less capacitance due to hit or miss audio caveats when the capacitance is too large.
I wanted to do something similar except avoiding the DC side, thus avoiding caveat.
So, far, I like the idea of a "snub the secondary" RC with a big enough resistor value to guarantee that it doesn't act like a cap. It would be nice to assure that whatever we did with the cap went into a known location--heating a resistor slightly. Based on the available data, I guess the resistor value to be somewhere in the range of 16 ohms to 96 ohms.
Can a resistor near that range be matched up with an X2 cap to create the needed RC?
Last edited:
Gootee, my point was that my resonance was at 14KHz, and a 1.5u film is still too small to effectively snub it without causing another 5KHz resonance. I mistakenly construed that the cap was the problem. And yes, a lytic could of course be used with the appropriate snubbing resistor, it's just I've been around DIYAudio too long and am lytic phobic
Again, capacitance does not snub.
For such a low-frequency resonance, you will probably just need a large-value resistor across the secondary, with no capacitor. If you could calculate the characteristic impedance, you could use the optimal R value.
If you know either the L or the C in the LC that is resonating, you can calculate the other one, because you know the resonant frequency.
Here is the usual procedure:
1. Measure the frequency of the resonance or ringing, using an
oscilloscope (or a circuit simulator, if you've modeled the parasitics well).
2. Add a shunt capacitor and adjust the value of this capacitor until the frequency of the ringing is reduced by a factor of two. I've left out the math but the value of this resulting capacitor will be three times (3X) the value of the parasitic capacitance that is creating the resonance.
3. Because the parasitic capacitance is now known, the parasitic inductance can be determined using the formula:
L = 1 / [(2 · Pi · F)² · C]
where F = (original) resonant frequency and C = parasitic capacitance.
4. Now that both the parasitic capacitance and inductance are known, the
characteristic impedance of the resonant circuit can be determined using the following formula:
Z = √(L/C)
where L = parasitic inductance and C = parasitic capacitance.
5. The resistor value used for the terminator or for the RC snubber circuit should be equal to Z, the value of the characteristic impedance, and the capacitor, if used, should be sized between four and ten times the parasitic capacitance. The use of larger (than 4X) capacitors slightly reduces the voltage overshoot at the expense of greater power dissipation in the resistor.
NOTE: The resistor, alone, is all that is needed to prevent or damp-out the ringing (or reflections, as the case may be). But if power dissipation in the R would then be too high, a C is added in series with the R, so that only the unwanted frequencies cause currents in the resistor. (And that is the only reason there's a capacitor in a snubber.)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.