I have also split the supplies to the SB3, standard factory switchmode supply the internal switchmode that generate 55v to the display and analog stages.
Super reg supply the internal regulator for 3.3V digital circuits.
Please see thread here for more info:
http://www.diyaudio.com/forums/powe...x-3-separate-powersupplies-2.html#post4273609
Super reg supply the internal regulator for 3.3V digital circuits.
Please see thread here for more info:
http://www.diyaudio.com/forums/powe...x-3-separate-powersupplies-2.html#post4273609
I've ordered 6x Panasonic TP series 25v 5100uf ESR 0.016 ripple 3480@ 125 degree C to replace the Unknown specified Panasonics in a psu of a commercial amp. Also some 35V 330uf of the same range for decoupling after the regulators.
http://docs-asia.electrocomponents.com/webdocs/0f80/0900766b80f8045d.pdf
Ha
http://docs-asia.electrocomponents.com/webdocs/0f80/0900766b80f8045d.pdf
Ha
RNS1E100MDS1JT | Nichicon 10μF 25 V dc Polymer Aluminium Electrolytic Capacitor, NS Series 2000h 6.3 (Dia.) x 7mm | Nichicon
Check these out for polymer, it's only 25v 10uf but the losses are 50uA
I wonder whether this can be a direct replacement for electrolytic at the same values without effecting electrical/electronic integrity?
Check these out for polymer, it's only 25v 10uf but the losses are 50uA
I wonder whether this can be a direct replacement for electrolytic at the same values without effecting electrical/electronic integrity?
Last edited:
Hi Music Lover,
Yes, they can. You do have to watch how you bend the leads as they are smaller than the Electrolytic parts. Just heed the max ripple current specs and other common sense things and you should be fine. Because these are new to the market still (as far as I'm concerned), I try to keep the actual working voltage down on them. I don't know how grumpy they might get over a brief over-voltage. I've heard they don't react well to reverse polarity, not even a little bit. So, just how much alike are these to solid Tantalum capacitors?
-Chris
Yes, they can. You do have to watch how you bend the leads as they are smaller than the Electrolytic parts. Just heed the max ripple current specs and other common sense things and you should be fine. Because these are new to the market still (as far as I'm concerned), I try to keep the actual working voltage down on them. I don't know how grumpy they might get over a brief over-voltage. I've heard they don't react well to reverse polarity, not even a little bit. So, just how much alike are these to solid Tantalum capacitors?
-Chris
Hi Welcome,
Personally, I think the topic is something that needs to be discussed. Something we don't often apply science to are the parts, and the reasons why one might be better than another depending on the circuit they are in.
There is a ton I don't know about how Nelson works, but I do know he experiments with parts to see how they react in different circuits. Isn't this the very definition of a scientific method??
-Chris
Personally, I think the topic is something that needs to be discussed. Something we don't often apply science to are the parts, and the reasons why one might be better than another depending on the circuit they are in.
There is a ton I don't know about how Nelson works, but I do know he experiments with parts to see how they react in different circuits. Isn't this the very definition of a scientific method??
-Chris
Reading his first post, he says:
Cool indeed!We can't always afford polypropylene capacitors, much less Teflon (tm), so we often make do with electrolytics when we need those higher values.
(...)
My favorite is the Elna silk capacitors...
(...)
There's hemp in them, too.
Reading his first post, he says:
Cool indeed!
Love it and I like Silmic II very much. Sometimes is Cerafine a better choice
I've come in very late on this thread, so please bear with me if I'm out of line. Someone like Nelson has the capability to test and measure all these caps and put them into real world circuits and get the feedback he needs. To start a thread like this, I believe he's just sniffing out what the general consensus is. Someone like me won't have that capability, all I have is an old CRT CRO that I scored free as it was not need anymore, and a Fluke 12 multimeter.
There is a general consensus in Alna Silmic II, but these come in small voltages and small capacitance, not suited to replacing 35v 3300uf. And I think Nelson is talking about using them in the signal path and not for psu or rail decoupling. I'm a great believer of measurements, to me the better the measurement the better the result, but not always.
@anatech,
I believe I read one of your post stating the higher the voltage capacity the better. If you look through the table of capacitor specs that seems to hold true for ESR. I'd like to try that, replace the same series of cap but of higher voltage and see what gains there are to the ear. I've just replaced an unknown Panasonic 22uf 35v to a known FC 22uf 50v which I can safely assume it's for rail decoupling as there are no caps in the signal path and notice a smoother SQ, and this is at its initial stage, no burn-in; but the increase smoothness can also be my head playing with me! New known caps form Unknown!!!
There is a general consensus in Alna Silmic II, but these come in small voltages and small capacitance, not suited to replacing 35v 3300uf. And I think Nelson is talking about using them in the signal path and not for psu or rail decoupling. I'm a great believer of measurements, to me the better the measurement the better the result, but not always.
@anatech,
I believe I read one of your post stating the higher the voltage capacity the better. If you look through the table of capacitor specs that seems to hold true for ESR. I'd like to try that, replace the same series of cap but of higher voltage and see what gains there are to the ear. I've just replaced an unknown Panasonic 22uf 35v to a known FC 22uf 50v which I can safely assume it's for rail decoupling as there are no caps in the signal path and notice a smoother SQ, and this is at its initial stage, no burn-in; but the increase smoothness can also be my head playing with me! New known caps form Unknown!!!
Hi Music Lover,
You can always try various capacitors in an existing circuit to see how they perform. Use the capacitor as a coupling cap. You can change the load resistance easily enough so you can hear an effect and see if the effect is better or worse with differing caps. Try to use a very low distortion circuit so it's problems don't swamp out the capacitor differences.
-Chris
You can always try various capacitors in an existing circuit to see how they perform. Use the capacitor as a coupling cap. You can change the load resistance easily enough so you can hear an effect and see if the effect is better or worse with differing caps. Try to use a very low distortion circuit so it's problems don't swamp out the capacitor differences.
-Chris
Hi Chris,
Believe it or not I'm changing these Caps in a low noise low distortion commercial preamp using LSK389 and a BB PGA2310 that's marketed $6000 AUD! I didn't buy this new, someone only had this a few months and it's been offloaded twice before I decided to check it out for a little over a 3rd of the RRP.
It's neatly laid out with outstanding built quality, not as liquid as the Manley tube amps but I'm a solid state boy! The pre has no inline caps for the signal path and all I have is psu smoothing and decoupling.
Hopefully swapping out the 6X 25V 3300uf with all 25v 5100uf won't over load the small 4 diodes that's used as a full wave bridge, I did order some 6A types from E14 but they are too big to be place on the PCB.
Hi Music Lover,
I guess I don't have to tell you my views are opposed to this. I've seen massive improvements by reducing the value of capacitance using improved capacitor technologies. Heck, if the unit in question uses too high values of capacitance, reducing it to normal levels also improves performance.
My views are also opposite those of Nelson's.
Yes, it is possible to have too high values of capacitance.
-Chris
I guess I don't have to tell you my views are opposed to this. I've seen massive improvements by reducing the value of capacitance using improved capacitor technologies. Heck, if the unit in question uses too high values of capacitance, reducing it to normal levels also improves performance.
My views are also opposite those of Nelson's.
Yes, it is possible to have too high values of capacitance.
-Chris
Frankly it's not the worst but it's far to be in the best lytic for decoupling and power supplies ! It has something cold despite a good resolution, the mid-highs are often too thin, without Flesch : hard to use it alone; tonal balance is never good and not natural at all! To picture: it sounds a little like plastic instead leather ! just my experience, mostly on the digital side and analogic side of sources like DAC & CD player mostly.
Last edited:
So, let me get this straight, you're findings are; more capacitance can deteriorate the SQ?
There's a lot of evidence to suggest this, so should I order 25v 3300 Panasonic TP series to try this theory out?
Hi Music Lover,
I would generally put something in that is close to the original if I haven't studied the circuit. I have been known to greatly reduce the amount of power supply capacitance in cases where the equipment designer has overdone things. The main thing to watch is the "hot switching current", which is the current peaks to charge the supply capacitance during normal operation. You can start to run into di/dt issues when the capacitance is too large. This is most easily seen with tube equipment and low power solid state, like preamps and CD players and the like. Power supplies can generally attenuate low frequencies much better than they can high frequencies. Once you create high frequency noise, it tends to travel throughout the product and is generally very difficult to get rid of. So some 120 Hz ripple is no big deal (especially if there is a voltage regulator circuit involved). Some added series resistance is helpful too.
For power amplifiers, the current needs are higher, but the same basic principles apply there too. There is a limit to supply capacitance where adding more doesn't really buy you much. In fact, adding more capacitance reduces the conduction angle of the rectifiers. That's generally a bad thing taken beyond a reasonable limit.
Look at your power supply DC rails with an oscilloscope. Look for nasty little peaks or ringing.
-Chris
I would generally put something in that is close to the original if I haven't studied the circuit. I have been known to greatly reduce the amount of power supply capacitance in cases where the equipment designer has overdone things. The main thing to watch is the "hot switching current", which is the current peaks to charge the supply capacitance during normal operation. You can start to run into di/dt issues when the capacitance is too large. This is most easily seen with tube equipment and low power solid state, like preamps and CD players and the like. Power supplies can generally attenuate low frequencies much better than they can high frequencies. Once you create high frequency noise, it tends to travel throughout the product and is generally very difficult to get rid of. So some 120 Hz ripple is no big deal (especially if there is a voltage regulator circuit involved). Some added series resistance is helpful too.
For power amplifiers, the current needs are higher, but the same basic principles apply there too. There is a limit to supply capacitance where adding more doesn't really buy you much. In fact, adding more capacitance reduces the conduction angle of the rectifiers. That's generally a bad thing taken beyond a reasonable limit.
Look at your power supply DC rails with an oscilloscope. Look for nasty little peaks or ringing.
-Chris
@anetech,
Chris,
The Panasonic 25v 5100uf TP arrived yesterday and I was able to put them in last night.
Stuck the CRO probe on the output of the preamp and there is zero DC. Didn't stick the probe on the rail voltage yet to look for spikes but because it was late, I decided to have a listen. Somehow in low levels at 22.15-23:45 just before shut eye, at low level listening there is a slight change in SQ, slightly more coherent, the background details are emphasised. Early days yet, but when I get back from work today, will definately stick the CRO on the rails and have a look.
In reference to pricing and availability, 3300uf 25v Isn't available in the TP range, the next available values are2700, 4700 and 5100uf and if you look at the E14 web site they are all available at $5.03 single quantity, same price.
EEUTP1E472 - PANASONIC ELECTRONIC COMPONENTS - Electrolytic Capacitor, TP Series, 4700 µF, ± 20%, 25 V, 18 mm, Radial Leaded | element14 Australia
EEUTP1V272 - PANASONIC ELECTRONIC COMPONENTS - Electrolytic Capacitor, TP Series, 2700 µF, ± 20%, 35 V, 18 mm, Radial Leaded | element14 Australia
There are a few amp designers that I have corresponded with and the have come to the same conclusion as you, it's the hot switching current that degrades the SQ. I believe that one way to resolved this is to have a bank of caps that's directly filtering after the bridge and then isolated that bank with RL with a secondary bank of caps that's closer to the voltage regulation with your main amp circuit.
Regards
Peter
Chris,
The Panasonic 25v 5100uf TP arrived yesterday and I was able to put them in last night.
Stuck the CRO probe on the output of the preamp and there is zero DC. Didn't stick the probe on the rail voltage yet to look for spikes but because it was late, I decided to have a listen. Somehow in low levels at 22.15-23:45 just before shut eye, at low level listening there is a slight change in SQ, slightly more coherent, the background details are emphasised. Early days yet, but when I get back from work today, will definately stick the CRO on the rails and have a look.
In reference to pricing and availability, 3300uf 25v Isn't available in the TP range, the next available values are2700, 4700 and 5100uf and if you look at the E14 web site they are all available at $5.03 single quantity, same price.
EEUTP1E472 - PANASONIC ELECTRONIC COMPONENTS - Electrolytic Capacitor, TP Series, 4700 µF, ± 20%, 25 V, 18 mm, Radial Leaded | element14 Australia
EEUTP1V272 - PANASONIC ELECTRONIC COMPONENTS - Electrolytic Capacitor, TP Series, 2700 µF, ± 20%, 35 V, 18 mm, Radial Leaded | element14 Australia
There are a few amp designers that I have corresponded with and the have come to the same conclusion as you, it's the hot switching current that degrades the SQ. I believe that one way to resolved this is to have a bank of caps that's directly filtering after the bridge and then isolated that bank with RL with a secondary bank of caps that's closer to the voltage regulation with your main amp circuit.
Regards
Peter
Hi Peter,
Going up a bit won't hurt things. It's the large increases in capacitance that breeds trouble. That is of course assuming that the starting values are in normal ranges.
I see where you're coming from by isolating the large capacitance bank, but you are still going to generate spikes. You may not see them on the scope as the dv/dt may be too high for the vertical amplifier, or even the probe. Once created, that genie is out of the bottle. It's best not to create problems to begin with. Also, your isolation will have impedance, and that means a resistive connection to power. You would be far better off leaving things as they were designed, using the first available value closest, or above, the original value.
The huge capacitor craze is much like every other "big thing" in audio. Remember tantalum capacitors? Never understood that fad. A little time and things righted themselves, but at what cost to customers? Same thing here, the sight of capacitors strung together and air wired to the circuit still can keep me up. They're still doing this!
Common sense. Works every single time.
-Chris
Going up a bit won't hurt things. It's the large increases in capacitance that breeds trouble. That is of course assuming that the starting values are in normal ranges.
I see where you're coming from by isolating the large capacitance bank, but you are still going to generate spikes. You may not see them on the scope as the dv/dt may be too high for the vertical amplifier, or even the probe. Once created, that genie is out of the bottle. It's best not to create problems to begin with. Also, your isolation will have impedance, and that means a resistive connection to power. You would be far better off leaving things as they were designed, using the first available value closest, or above, the original value.
The huge capacitor craze is much like every other "big thing" in audio. Remember tantalum capacitors? Never understood that fad. A little time and things righted themselves, but at what cost to customers? Same thing here, the sight of capacitors strung together and air wired to the circuit still can keep me up. They're still doing this!
Common sense. Works every single time.
-Chris
Hi Chris,
The Panasonic 35v 330uf TP series arrived late yesterday evening. These are the same values I've decided to replaced the known Panasonic NHG 25v 330uf. Notice I ordered the exact replacement values apart from voltage ratings, this is because these are after the standard 7812/7912 regulators.
I also checked the transformer secondary with the CRO and as expected the clipping you see on the top and bottom waveforms has expanded telling me the diodes are now working possible twice as hard as before. The turn off spike is also there, the DC supply rails are have clean traces, or as clean as I can see with this 60mhz CRO. And the power consumption meter shows that it's increased from 5.7W to 6.8W.
The line output is also clean and free of any DC so that was the green light to put in the rig and have a listen.
The SQ changed from an airy quality where there is space around instruments and voices that was almost a holographic presentation. I can relate to why some on these neck of the woods post that it's analytical and lifeless. This was after I placed the 5100uf x8. Now that the 330uf are in place it's brought back a slightly more lifelike presentation and I also have front to back depth. Early days yet as I only had an hr of listening before shut eye.
Just to add to my previous post, I've seen some designers placing supply filter caps on the PCB very close to the output devices in power amps. I can see tha advantages there.
Peter
The Panasonic 35v 330uf TP series arrived late yesterday evening. These are the same values I've decided to replaced the known Panasonic NHG 25v 330uf. Notice I ordered the exact replacement values apart from voltage ratings, this is because these are after the standard 7812/7912 regulators.
I also checked the transformer secondary with the CRO and as expected the clipping you see on the top and bottom waveforms has expanded telling me the diodes are now working possible twice as hard as before. The turn off spike is also there, the DC supply rails are have clean traces, or as clean as I can see with this 60mhz CRO. And the power consumption meter shows that it's increased from 5.7W to 6.8W.
The line output is also clean and free of any DC so that was the green light to put in the rig and have a listen.
The SQ changed from an airy quality where there is space around instruments and voices that was almost a holographic presentation. I can relate to why some on these neck of the woods post that it's analytical and lifeless. This was after I placed the 5100uf x8. Now that the 330uf are in place it's brought back a slightly more lifelike presentation and I also have front to back depth. Early days yet as I only had an hr of listening before shut eye.
Just to add to my previous post, I've seen some designers placing supply filter caps on the PCB very close to the output devices in power amps. I can see tha advantages there.
Peter