EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: LinuxHata on July 06, 2023, 05:29:34 am
-
Hello. I have a following situation.
A DC load should be powered from AC mains via the transformer (no SMPS).
It requires 30+-2 volts at 5 amps.
I can get matching transformer, but there is an issue.
Say, transformer outputs 30V, after bridge rectifier and capacitors, at no load there will be about 41VDC, which is too high for the equipment. If I select transformer in a manner, that it provides 30V under no load, it should output (30+0,6)/1.41=21.7 volts. But, under the load, output of bridge rectifier will sag from 30V to that 21V, and equipment will turn off.
So how to fix this?
Add linear pre-regulator to 41VDC output?
-
If transformer is appropriate then DC voltage ratio to AC under max load is somewhere 1.25 of secondary AC voltage. (30/1.25 = 24 V).
So, you need either 24VAC 7-8-10 A rated transformer (not 5 A rated), or just some DC/DC converter for you existing 30VAC/42VDC transformer (if your load is fine with high frequency pulsations instead of low-frequency pulsation. Typically that's ok).
-
No easy answers there.
Actually answer is: use SMPS.
Why?
1) Classical PSU (with transformer) output voltage is heavilly dependant upon output current. As you calculated + resistance drops.
2) Classical PSU output is proportional to mains voltage. If mains will drop 10% output will drop 10%
3) Linear post-regulators will be hot. For most regulators for correct operation about 3 V drop is needed -> 15 W loss at low line in most optymistic case. About 50 W in less favourable cases.
4) Going international 100-240 VAC 50/60 Hz is huge pain. Probably not your case.
Solutions:
1) Carefully design your transformer PSU. Not easy to find stiff enough transformer, and will be bigger/heavier/more expensive.
2) Widen circuit allowable voltage say 30 V +/- 5 V
3) Add heavy and hot linear regulator + accept power waste
4) Use SMPS. Quite a number of 24 VDC PSUs can be adjusted to 28 V
-
Add a regulator. Linear power supply is highly inefficient, there is no way around it. Use an SMPS instead.
-
What is the current draw at "no load"?
Using linear regulator isn't that big of a problem if your load only ever work in two state (i.e. fully on with lots of current draw, or fully off/standby/sleep/idle with little current draw).
-
A perenial issue. Its not easy to get a stable output invariant with current draw and secondary voltages. Well, bloody impossible to optimise.
I always avoided smps in audio and instrumetation projects. They got better with time and I got better at applying them. Now I am a believer.
Regulation, transient response, in band ripple and noise are so much better.
With a good mains and output filter, good ground wiring, shielding and some faith, it will all come good.
-
You require +/-6.7% accuracy of the supply voltage, this is impossible to meet across all variables without voltage regulation.
Or (once again) SMPS.
-
All I need to do, is to clamp output voltage below 32VDC with no load (around 5mA consumption) condition.
I do not need it to be stabilized, the load will take care of itself, just output voltage should not exceed 32VDC and would not sag below 28V under 5A load. The idea of "big" transformer sounds practical - I have some 24V, 15A transformers lying around. I guess, I can add or reduce number of secondary turns to get the required voltage.
Another idea which just came to my mind is to have multi-tap transformer, and use relays to switch windings in that way, that output voltage in idle (no load) mode won't exceed 32VDC.
Actually, this can be done with 2 relays, zener and transformer which has dual 110V coils.
relay is connected to DC output via the zener and some resistor, for current limiting.
In turned off state, the transformer 110V winding is connected thru the relay to ac mains.
As power is being applied, voltage on output capacitors rises, and at certain point, reaches the value when zener conducts (zener selected in that way, that conduction occurs when DC output exceeds 32VDC), causing relay to switch, and connect 2nd 110V winding in series with first one, causing output voltage to drop below 32V. Capacitor will start to slowly discharge via the relay, and once voltage drops below of that of zener opening voltage, relay releases, causing winding switching and voltage increase again. This will result in a clicking sounds and some voltage fluctuations, which can be a bit annoying.
Above was "no load" condition. When we have the load, output is loaded, so it never exceeds 32VDC and relay never triggers and everything is ok.
What do you think?
-
You may also take into account the ripple from the capacitors
(about 4V if you use the rule of thumb 2000uF/A).
-
A depletion MOS follower and zener diode may do what you want. Such a voltage limiter is sometimes used for automotive transient protection.
A shunt regulator isn't a good idea because it will get overloaded by high-line conditions. Maybe rare, maybe out of spec so it's technically the power company's fault, but I'd rather have the pass limiter.
Alternately, get a better transformer, one that's not so saggy. Typically this behavior occurs with small split-bobbin type transformers (often small enough to be "impedance protected" i.e. they simply can't deliver enough short-circuit current to catch fire), but I suppose it might happen with any size of the same design. Shell (E-core, windings in the same place) and toroid types have excellent regulation. You also need enough filter capacitance that the output isn't drooping severely under full load (typically DC output 1.2-1.3 times AC Vrms).
Tim
-
You could use something like this SCR pre-regulator as shown in figure 5 (https://www.analog.com/media/en/technical-documentation/application-notes/an32f.pdf) of AN32-3.
-
Alternately, get a better transformer, one that's not so saggy.
output voltage should not exceed 32VDC and would not sag below 28V under 5A load. The idea of "big" transformer sounds practical
With those numbers I think even a very 'stiff' transformer might struggle. Vf on the bridge will go up and sag on the filter capacitors will increase. Even with very large filter caps, you'll still have a very small conduction angle and thus a very high RMS/AVG current ratio.
-
That SCR pre-regulator looks really interesting, thanks!
But my transformers are not weak - For example, I have one which has secondary rated at 20V 5A. When I change load on it from 0A to 5A, AC voltage goes down from 20.7 to 20.1 volts. But the rectified voltage drops from 28V no load to 21 volts at full load. So I don't think that installing larger transformer will help. I also tried to increase output capacitor value from 2200uf currently to 10000uf - zero difference.
-
You also need to account for variations in mains voltage - for example, here at 2-4 AM the voltage may go up to 240-245v AC which is acceptable because power grid is 230v +10% , -5% or something like that.
If you ignore that, I suppose you could get a 24v AC transformer, and use two or three bridge rectifiers in series to drop a bit of voltage. For example put a heatsink between two GBU806 bridge rectifiers - https://www.digikey.com/en/products/detail/diodes-incorporated/GBU806/278614 (https://www.digikey.com/en/products/detail/diodes-incorporated/GBU806/278614) - and connect the output of first to input of the second and now your bridge rectifier will drop around 4v
So 24v x 1.414 - 2 x 0.8v - 2 x 0.8v = 30.5v - ish ... with higher voltage on mains, could go to 32v and still be fine.
Add some capacitance, like maybe 15000uF + , to smooth out the output of transformer.
(C = 5A / (2 x 60 Hz x (30.5v - 28v) ) = 5/300 = 0.016666 = 16,666 uF
-
Ah, you're overloading your transformer also. 5A AC RMS rating, at a typical power factor of 0.5 for capacitor-input rectifier, is only 2.5A output. But that's 2.5A at elevated voltage, so it's not a complete waste.
The thing about leakage inductance is it's not necessarily a (50/60Hz) regulation issue. The rectifier draws harmonic currents, which are weighted more heavily by leakage inductance, thus the output can droop more with a rectifier.
Could you enlighten us what type transformers they are?
Typical capacitance is in the 1-2mF range for a 12V 1A supply. So, let me see here, 30V can use 30V/12V times less capacitance, but needs 5A/1A times more as well, or 2-4mF overall.
Tim
-
That SCR pre-regulator looks really interesting, thanks.
But it's ancient and crappy form of switch-mode converter, really, with all the downsides and nearly none of the benefits. What is your true reason for ruling SMPS out?
If you want to build the thing yourself and are afraid of the complexity/dangers of designing a mains-powered SMPS, then a viable option is to keep your transformer + rectifier + capacitors but use a simple buck converter to convert to lower DC voltage, with much better efficiency than linear regulators.
-
All I need to do, is to clamp output voltage below 32VDC with no load (around 5mA consumption) condition.
I do not need it to be stabilized, the load will take care of itself, just output voltage should not exceed 32VDC and would not sag below 28V under 5A load. The idea of "big" transformer sounds practical - I have some 24V, 15A transformers lying around. I guess, I can add or reduce number of secondary turns to get the required voltage.
Another idea which just came to my mind is to have multi-tap transformer, and use relays to switch windings in that way, that output voltage in idle (no load) mode won't exceed 32VDC.
So what you are describing IS a regulated power supply specified for an output of 30V with a tolerance of +/- 2V
For a transformer+rectifier the output will always vary with the mains input voltage and output load. So the only way to make it stable like that is some sort of regulated feedback.
Most of the methods of doing this are going to produce heat or be complicated. The easiest way to do this is to use a switchmode DC/DC converter on the output to regulate DC output of the rectifier. But once you have a DC/DC it is easier to just simply use one that can take 220V AC as input and directly give you the DC. This will make the whole power supply smaller, lighter and produce even less heat than the transformer alone.
A PSU brick module that does 30V 5A output costs only 26 USD including shipping:
https://www.aliexpress.com/item/1005002282450541.html (https://www.aliexpress.com/item/1005002282450541.html)
(Not that i would recommend using the cheapest Chinese module, but you can get a proper namebrand MeanWell PSU for not that much more)
Even if i had a suitable transformer it would cost me as much in parts to get that transformers output regulated, let alone if i factor in the cost of my time to figure it out and put it together.
-
So what you are describing IS a regulated power supply specified for an output of 30V with a tolerance of +/- 2V
This, "I don't need regulation, I just need regulation to 6%". Any random regulated power supply circuit would be something like 3% accurate because that's what you get with lowest cost bandgap references built-in regulator ICs plus 1% resistors etc. There probably is nothing simpler available for the 3%-6% range acceptable to the OP.
A (genuine) MeanWell power supply is the most obvious solution. If this is going to be a power supply learning experience, then I do suggest the mains transformer + bridge rectifier + buck converter approach; the low-voltage DC buck is an order of magnitude simpler (and safer!) to design than a mains SMPS, but still a suitable level challenge.
Esoteric ideas like AC thyristor switching will be disappointing in performance and as difficult if not more difficult to engineer than better, more usual solutions.
-
I need a simple thing.
A DC power supply which will be outputting exactly same DC voltage as AC voltage entering it.
I need to remove that 1.41 multiplier somehow.
That's all.
-
If you are not too worried about the efficiency, then just add a 1A bleeder load to the output of your rectifier. That should be enough load to provide a decent voltage with no additional load.
-
I need a simple thing.
A DC power supply which will be outputting exactly same DC voltage as AC voltage entering it.
I need to remove that 1.41 multiplier somehow.
That's all.
:palm:
At least you didn't ask that it also have the same AC input current as the DC output current!
Seriously, the problem is that the fundamental nature of AC voltage is different than the DC voltage you wish for. The 'equivalency' of a given AC (RMS) voltage and a DC voltage is only that if the two are the same, they will heat up a resistor by the same amount. Other than that, they are two different animals.
-
I would also want a very simple converter which takes 100W of input power and outputs 200W. See? Requirements are easy to come up. What matters is if that is achievable given physics, available components, and human knowledge of existing circuits. Your requirements specify a completely normal, bog standard mains -> DC voltage power supply, gazillions are manufactured every day. If there was some simpler way to achieve that, it would be already in use. (Your looser 6% regulation spec, over the usual 2-3% of cheap regulated supplies, unfortunately does not offer any extra simplicity I could think of.)
AC is not magic. Just look at the sinusoidal signal and understand that to output a steady DC voltage, you need to store the energy somewhere, and to output a steady voltage within your min/max range, you need to convert voltages to different ones. This is essentially what a SMPS does.
-
I'm worried about efficiency, this is why I asked :) I don't want to use bleeder resistors or whatsoever kind of "shunt" regulator.
I know how things work and I know the physics. I need to accomplish a specific task, and I came to switchable winding idea.
I think I'll give it a try these days and let you know. If it doesn't work, I'll go with thyristor control way.
-
All I need to do, is to clamp output voltage below 32VDC with no load (around 5mA consumption) condition.
I do not need it to be stabilized, the load will take care of itself, just output voltage should not exceed 32VDC and would not sag below 28V under 5A load.
Another option is to use a LDO linear regulator and set it for 32V.
Then when current draw is low, the LDO keeps it's output voltage at 32V but doesn't dissipate too much power because there's not much current passing through it.
Then when the load is higher (say 5A), and the transformer DC rectified voltage sags down to 29V, the LDO is fully "on" and only dropping around 1V, so it'll be dissipating 5 watts.
-
I'm worried about efficiency....
...I know how things work and I know the physics.
I need to accomplish a specific task, and I came to switchable winding idea.
If you aren't doing this as an learning exercise, can you just briefly explain why you've rejected the COTS SMPS path?
https://www.trcelectronics.com/View/Mean-Well/EPP-200-27.shtml (https://www.trcelectronics.com/View/Mean-Well/EPP-200-27.shtml)
Linear (transformer/rectifer/capacitor) single-phase AC-to-DC power supplies typically have efficiencies in the 50% range and always have the issue of voltage sag no matter how robust the transformer. This is an inescapable result of, among other things, the fact that this design only uses a small part of the AC waveform and the majority of the time it doesn't conduct. At low loads the conduction angle (on-time) is very, very low and the DC voltage will be very near the peak AC. At higher loads, the conduction angle increases and the maximum available voltage decreases. You can't win.
This problem has presented itself to designers for many decades and no simple solution has arisen. Switchable windings are used in a few cases--I have a linear lab PSU and audio amplifiers that use this method to reduce heat loss--but they are huge, expensive devices that don't even achieve 50% efficiency. Thyristor control has been done in some cases, but it is barbaric, noisy (EMI and usually audibly) and has a terrible power factor. Again, if you are experimenting for learning purposes--go for it. But if you just want a good solution, do what everyone else does--find a solution that someone else came up with and copy adapt it. And if you can't find such a solution, in the words of the great Arsenio Hall that should "make you say Hmmmmmm."
-
Yes this is "single-use" project for own purposes.
I do not want to use SMPS for various reasons in this project, and in general, I don't like SMPS at all. All my household appliances, where possible, run either from transformer based supplies, or I have all SMPS supplies modified, by removing Y capacitors and breaking the middle point connection of input AC filter with output GND terminal.
-
If you remove Y capacitors from SMPS noise rises significantly... What is reason for Y caps removal?
Today I would rather filter SMPS than play with classical PSU. It is only a preference however.
-
You would be surprised how many SMPS power supplies there are around you.
Pretty much any commercial product that needs low voltage DC somewhere inside it (this is almost everything due to MCUs running the show in most appliances) also has a SMPS converter inside of it. This means TVs, radios, microwaves, computers, battery chargers, washing machines, ovens, stoves, toasters, air conditioners desk lamps, alarm clocks, most kinds of LED lights... etc. Even traditionally dumb motor appliances like hair dryers, drills, saws..etc are now run with switching converters because they now use BLDC motors.
I don't think i even seen any brand new consumer product with a iron transformer inside for the last 10 years. Even test equipment is all going to switchmode, things like oscilloscopes, spectrum analyzers, signal generators, bench DMMs...etc are all now powered by SMPS. Despite it being important for test equipment to not be electrically noisy, but that is because they are well designed switching converters so they make very little noise.
All this is for a good reason. Power supplies with oldschool iron transformers or linear regulation are big, heavy, cost a lot of material to make then, yet after all of that they are very inefficient and make a lot of heat.
Yes there are a lot of crappy badly designed SMPS converters out there. but the well designed ones are very efficient, low noise and robust.
Also why would you remove the Y capacitors out of power supplies. That just makes them spew electrical noise out of the PSU.
-
Some months ago we got two new linear lab power supplies, brand is OWON. They use transformer secondary taps with relays to adjust the rectifier output voltage to what is required.
Or get a 5 A capable PWM module to use it as a secondary switcher. € 5 .. 10 at ebay.
Also a 24 V 5A switcher is small and cheap, like Meanwell NDR-120-24.
Regards, Dieter
-
Yes this is "single-use" project for own purposes.
I do not want to use SMPS for various reasons in this project, and in general, I don't like SMPS at all.
Is there some reason for this? You realize the thyristor circuit you are looking at is an SMPS, and an extra crappy one?
-
You forget how many re-inventors of the wheel lurk on this blog.
Why build what you can buy- Because we love pain and expense. Masochistic tendencies/ OCD.
I want to learn, but not from the mistakes of others. Narcissistic personality disorder/ RTFM
I feel bad about taking the path of least resistance. Martyr complex and closet Cap reformer.
I dont like X. Yeah, but why. I just don't you B**ta**s. Do I need a God damn reason?? Get off my case!! Anger issues/ Alcohol Abuse /sales and marketing
I want to own common knowledge as my own discovery. Low self esteem/ Murderer in the making/ Rail gun fan boy.
I am unable to trust, even f**kin* datasheets are lying to me. Naked Egomania /Incipent Pyschosis/ Ebay Opamps
There must be an Arduino, Linux and 555 based solution, its on GitHub surely?? Delusional/ Utterly Delusional.
I want to make an audio preamp using a thyristor, any advice. Jeez, again?
Pedantry is a given.
-
I do not want to use SMPS for various reasons in this project, and in general, I don't like SMPS at all. All my household appliances, where possible, run either from transformer based supplies, or I have all SMPS supplies modified, by removing Y capacitors and breaking the middle point connection of input AC filter with output GND terminal.
I assume you are concerned about leakage currents then. It's true that there are applications where a linear PSU is appropriate, but you have to live with their inherent characteristics. I power the LED lights on my bench with a linear supply that I made up myself with the main goals of reliability and low noise (the originals were terrible), power factor and efficiency being secondary. I ended up with extremely low noise--lower than any SMPS--but an efficiency of 55% and a PF of 0.81. It would be very difficult to signficantly improve on that.
-
Medical grade very low leakage current SMPS are easily available and do not even cost arm and leg. I remember using such from Meanwell since it happened to be the most affordable option in certain voltage/current at that time, even though I didn't need that low leakage. Such supplies are two-prong and touching the outputs while being grounded you feel basically nothing. They still might have an Y cap there but it's very small value.
Just buy a decent $35 supply instead of the $25 Ebay copy.
-
“ Martyr complex and closet Cap reformer.” :-DD
The best phrase in this whole thread.
-
Yes, I don't like equipment "biting" me when I touch it and something grounded same time, or, if I'm charged with static, I touch the device and discharge goes thru that y-capacitor/filter caps path and I'm getting a shock. For same reason, I cut the grounding line of my soldering iron (it's handle and wire are anti-static, so they are conductive), and feeding it thru isolating transformer.
For the thyristor circuit, yes I understand that it is some kind of "flintstones" era SMPS, but it has no Y capacitors and switching frequency is fairly low, removing almost all issues currently active with "normal" SMPS.
-
A linear regulator won't be as wasteful as you might think, when the load is applied the secondary voltage will drop, eventually the regulator will be almost a voltage follower with a small 0.7V voltage drop (Because it's Darlington), wasting at most 5W at 5A load (150W).
(https://www.circuitstoday.com/wp-content/uploads/2011/10/regulated-dc-power-supply.png)
-
Thanks! I also thought that way, but with LM338K or similar. I was hoping that might there be some chance to reduce power waste further, but if there's no other way, I'll do that.
-
Yes, I don't like equipment "biting" me
So just solve this problem. There are two ways, use earthed (three-prong) supplies, or get low-leakage type supplies.
For the thyristor circuit, yes I understand that it is some kind of "flintstones" era SMPS, but it has no Y capacitors and switching frequency is fairly low, removing almost all issues currently active with "normal" SMPS.
You realize the low f_sw only makes it worse? Switch mode power supplies have been getting better thanks to going to higher frequencies, which are easier to filter with smaller and cheaper filtering components.
Also the DIY thyristor controlled thing will not pass EMI regulations, so it's totally apples-to-oranges comparison.
-
How many cell phone chargers you have seen with 3 prongs? :D
I've only seen ones with UK plug, and that 3rd prong is plastic anyways....
-
I don't understand why that is relevant to the discussion. Are you now building a cell phone charger?
Your 30V 5A supply could obviously be earthed, even if cell phone chargers aren't. In fact, most if not nearly all off-the-shelf SMPS supplies in that power scale are earthed, 3-prong, so your "problem" is completely made-up.
-
When something is earthed, and you have the static potential on you, when you touch the earthed object, that potential discharges thru it, and you get a bite. I don't want to get a bite :) So this is I do not want to earth anything :)
-
Bad attitude. Other people using electronics want the safety. In general it's required by law anyway.
You should rather try and get rid of your ESD generator. There are ways to solve that problem and that's the recommended procedure. Also better for electronics.
Regards, Dieter
-
When something is earthed, and you have the static potential on you, when you touch the earthed object, that potential discharges thru it, and you get a bite. I don't want to get a bite :) So this is I do not want to earth anything :)
You have that exactly the wrong way, too! Those static charges accumulate exactly because earthing is insufficient. By trying to make things even more floating and isolated, you are making the situation worse and worse. It's like this saying that peeing in your pants helps keep you warm in winter - but only for a while.
The real problem is, I guess, something like carpets made of plastic fibers, combined with dry air in winter. Removing earth connections from electric appliances won't solve the root issue at all. You can't stop earthing certain classes of devices, as this is mandated by law for safety (or, you would have to do massive rebuilds to convert them into double insulated).
-
I did not read all of the above.
The simple answer is to use a transformer that is rated for 2X or 4X or 10X the maximum current your circuit will draw. That way, it will have a lot smaller internal resistance and will not drop as much Voltage under load.
No loss in efficiency. No regulator, switching or linear. Just a much more expensive transformer.
I can do it cheap. I can do it well. I can do it efficiently. PICK ANY TWO.
-
I find it really rare for 2 prong switching power supplies to have so much AC leakage that you can actually feel it.
For example i measured a genuine Samsung phone charger model EP-TA200. Like most it is two pronged and even on our more spicy 230V mains it clocked in at about 7uA RMS of leakage to mains earth. That is basically nothing, no way you could feel that since you need more like 0.5mA to clearly feel it (or perhaps you can feel down to 0.1mA if you find just the right conditions, like you contract it at a sharp point to concentrate all that current trough a sensitive part of your finger)
However when it comes to electrostatic shocks. Not having the capacitors there will NOT help at all. Those discharges are in the 10s of kV and the insulation on floating power supplies is not capable of holding it back, it will just arc over somewhere. In fact this can actually be DANGEROUS! A small arc jumping over to a low voltage and high current power source can actually start conducing the high current across. This is how high frequency arc starters on TIG welders work, very short high voltage pulses cause tiny arcs to jump across to open up a channel for the welding current (only 10s of volts at 100s of amps) to flow and arc, without having the electrode touch the metal. By removing safety components from a SMPS you could create a risk of this same thing happening inside the power supply. The static discharge might arc over at a point where live mains is present, giving it a path to arc across into the isolated DC side. If you are touching the isolated side, this means the arcing current flows trough your body. Fair enough that this being AC the arc likely get extinguished quickly, but can still damage the insulation in the PSU, potentially causing it to develop a isolation fault.
The way to combat ESD is to address the sources of electrostatic charge that make your body charge up. This could be the materials of your clothes or slippers/shoes (synthetic, wool..etc), could be the materials of your floor (like carpet), could be furniture. Air humidity has a huge effect on all of these. Having >50% humidity tends to make static buildup very difficult to happen, because of this it only really happens where i live in the winter time. So you could consider just using some air humidifiers to raise it. There are also some sprays that you can use on the floor to make it easier for the charge to bleed away (i hear even just diluted fabric softener helps a lot).
-
I was assuming you were concerned about high frequency noise until you mention SCR pre-regulation. That's just SMPS but worse....
If you just want completely floating supply and/or isolation then surely using actual isolation transformer beefy enough for the whole mains should be a better solution right? No need to individually "modify" all of your power supplies.
Yes, I don't like equipment "biting" me when I touch it and something grounded same time, or, if I'm charged with static, I touch the device and discharge goes thru that y-capacitor/filter caps path and I'm getting a shock. For same reason, I cut the grounding line of my soldering iron (it's handle and wire are anti-static, so they are conductive), and feeding it thru isolating transformer.
You're "solving" what looks like an already bad mains wiring and grounding in your house by jeopardizing your safety even more. Great...
Ask an electrician to check your mains wiring. Ask them too if it's okay (and allowed within code) to convert your mains to TN-S or IT grounding system if it wasn't already. Don't hurt yourself please.
-
So you could consider just using some air humidifiers to raise it. There are also some sprays that you can use on the floor to make it easier for the charge to bleed away (i hear even just diluted fabric softener helps a lot).
Yup. If all the OP is concerned about is getting zapped with static electricity, then this is the way.
No power supplies or grounding involved, but the worst source of static zaps involve getting into my car in the winter. The car is definitely NOT grounded because it's sitting on rubber tires and, since it is gasoline powered, it is not connected to the electric grid at all. Yet I get zapped when I touch it in a low humidity environment. Never happens during the rainy season or summer when the humidity is higher.
-
The simple answer is to use a transformer that is rated for 2X or 4X or 10X the maximum current your circuit will draw. That way, it will have a lot smaller internal resistance and will not drop as much Voltage under load.
That will help, but unfortunately not that much, because the core issue still remains, namely the voltage ripple from rectified sine wave. But because OP currently has a seriously undersized transformer, the first step would be to try what you suggest and see if it is enough.
-
I don't think i even seen any brand new consumer product with a iron transformer inside for the last 10 years. Even test equipment is all going to switchmode, things like oscilloscopes, spectrum analyzers, signal generators, bench DMMs...etc are all now powered by SMPS. Despite it being important for test equipment to not be electrically noisy, but that is because they are well designed switching converters so they make very little noise.
Now iron transformers go only to very low end, low power products and high end, low power (20 W?) products like 6 1/2 digit DMMs. This means classic PSU makes sense only when ultra low noise is required. Rare case in general.
My issue with OP attitude is removal of Y caps does not help with electrostics (actually it may worsen ESD effect), surely worsens the EMC emissions as well as suggests issues with home (?) electric installation.
Also ESD problems can be really solved with removal of source, not victim of discharges. With even 15 kV discharges using normal clothing preventing zap from charged body is virtually impossible. Better prevent charging.
-
Pretty sure OP's problems are caused by poor PE wiring. It's either non-existent, suffering high earth impedance, or combined with neutral. The "zapping" is from the two Y caps forming a voltage divider at half mains voltage to the appliance case. It almost certainly have nothing to do with actual statics.
-
Well, I guess a lot of you have standard Apple USB charger, two prongs, yes.
Connect it to mains, plug in cable, hold the contacts by one hand and touch some grounded object, like steel faucet, with another. Enjoy. And no, my apple charger is not fake, it is genuine one, which came with iPad. Same for other chargers. And I really don't understand, how grounding wire can help in case with 2 prong chargers leaking ac current to output via the filter and Y capacitors.
Changing humidity is not good idea - why change my living habits, introduce mold and moisture, when I can simply just remove pair of capacitors?
For the transformers, as I said, there is no sag at AC side at full load. The transformer I'm currently using comes from 70s and is rated at 15 amps 24 VAC secondary. The DC voltage is thing that varies.
By the way, I solved the issue, but not in a perfect way. Output of rectifier is connected to a bleeding resistor via the zener diode. That's all. This limits the top voltage. As load arrives, voltage drops, zener stops conducting and everything is fine.
-
standard Apple USB charger, two prongs (-) Connect it to mains, plug in cable, hold the contacts by one hand and touch some grounded object, like steel faucet, with another. Enjoy. And no, my apple charger is not fake, it is genuine one, which came with iPad. Same for other chargers.
If I remember regulations correctly for Class II devices leakage current is 250 uA max (240 VAC 60 Hz) It is beyond detection level for humans. Normal humans however. Very moist skin, certain deseases increase sensitivity. Still I doubt 250 uA can be "upleasant feeling".
Charging current even for 2.2 nF .. 2.7 nF capacitor seems strage it can "bite". Is instalation correct? Are leakage currents low enough? Maybe there is something else which leaks or pumps spikes of higher voltage which induces "biting" senseations?
Actually, if all Y caps are removed across house, including those in EMC filters, then there can be significant amount of HF currents circulating around, and those could at least potentialy "bite".
-
You are correct, the steady state (mains frequency) AC current leakage through the Y caps is very low. There is another aspect though, the instantaneous (fast rise-time) charge equalisation when you make contact between the secondary side and ground, either directly or via your body. This rapid discharge is significantly higher current than the steady state mains frequency leakage and can be felt. It is a one time event though (unless you keep disconnecting and reconnecting) its magnitude also depends on where in the mains cycle you make the contact and the series (body) resistance.
The current through a capacitor is always dependent on frequency / rise time. As you say, it is very low at mains frequency, and the capacitor values, yes around 2n2, are sized so that they can never pass a dangerous current under such event's. Their action is required though, a couple of hundred uA is never going to do anything useful to suppress EMC at RF - their impedance is very low at the frequencies that they are there to supress. The minor inconvenience of a small (and safe) instantaneous zap if you touch the connector and a solid ground at the same time is inevitable.
@LinuxHata:
Removing the Y caps just blasts the world with EMC - Without them you will have significant RF current flowing through you (much higher than the normal Y cap leakage), the frequency is just too high for you to feel, it doesn't mean it isn't there though!
-
You're needs voltage regulator on the rectifier output. You have two options:
1) Linear regulator - it has clean output and low noise which is good for sensitive equipment, but it has energy loss on heating
2) Switching mode regulator - it has dirty and noisy output which will be serious issue for sensitive equipment, but it has low energy loss on heating
-
And I really don't understand, how grounding wire can help in case with 2 prong chargers leaking ac current to output via the filter and Y capacitors.
With the third prong, the Y capacitor is connected to earth directly. The exact same earth you are touching. They are at the same potential, hence no zap at all. The small leakage current isn't flowing through your body, it's flowing through the earth wire.
If you still feel a zap with 3-prong SMPS connected to a 3-prong socket, your house wiring is seriously and dangerously broken and potentially lethal, and needs to be fixed by an electrician. Or have you modified your house wiring too, "to stop zaps"?
Changing humidity is not good idea - why change my living habits, introduce mold and moisture
You don't need to go to mold-growing levels. Very dry air isn't good to your respiratory system or skin, anyway. This is the primary reason to humidify, to get from that horrible 10% "zaps and skin problems" range to some manageable 30% region.
when I can simply just remove pair of capacitors?
A) because those capacitors have absolutely nothing to do with your static electricity problems, and removing them thus does absolutely nothing to it; you are seeing a placebo effect or something like that.
B) Because the capacitors are there for a good reason, and you are modifying your equipment and making it illegal, possibly causing problems to your other devices or your neighbors. Do you also drive drunk because "muh it's my habit"?
I would suggest you stop modifying and building mains-powered devices and keep to SELV stuff. You clearly don't have enough understanding to do that safely, and your attitude problem makes it much worse than just lack of understanding alone.
-
standard Apple USB charger, two prongs (-) Connect it to mains, plug in cable, hold the contacts by one hand and touch some grounded object, like steel faucet, with another. Enjoy. And no, my apple charger is not fake, it is genuine one, which came with iPad. Same for other chargers.
If I remember regulations correctly for Class II devices leakage current is 250 uA max (240 VAC 60 Hz) It is beyond detection level for humans. Normal humans however. Very moist skin, certain deseases increase sensitivity. Still I doubt 250 uA can be "upleasant feeling".
Charging current even for 2.2 nF .. 2.7 nF capacitor seems strage it can "bite". Is instalation correct? Are leakage currents low enough? Maybe there is something else which leaks or pumps spikes of higher voltage which induces "biting" senseations?
Actually, if all Y caps are removed across house, including those in EMC filters, then there can be significant amount of HF currents circulating around, and those could at least potentialy "bite".
One possibility is that the Earth wire is broken somewhere in / after the panel, so that the sockets have one continuous net of "earth" and then there is another "earth" net connected to the physical earth, plumbing etc. In such case, all Y capacitors of all grounded switch mode power supplies act in parallel for significant total leakage current. And, any device with actual hard ground fault becomes a death trap.
-
By the way, I solved the issue, but not in a perfect way. Output of rectifier is connected to a bleeding resistor via the zener diode. That's all. This limits the top voltage. As load arrives, voltage drops, zener stops conducting and everything is fine.
I was going to suggest this, but you'd said to someone else that bleeder resistors were out due to efficiency requirements.
The only problem with that method (depending on zener voltage and resistor value) is that small variations in output voltage (Due to line voltage changes) will result is large changes in bleeder/zener current.
As long as the components are rated to handle the power, at the max possible line voltage input, then you should be good to go. What components did you use?
Another method I've seen used is to sense the load current and then switch the bleeder resistor on/off accordingly. (With hysteresis of course)
-
Well, I guess a lot of you have standard Apple USB charger, two prongs, yes.
Connect it to mains, plug in cable, hold the contacts by one hand and touch some grounded object, like steel faucet, with another. Enjoy. And no, my apple charger is not fake, it is genuine one, which came with iPad. Same for other chargers. And I really don't understand, how grounding wire can help in case with 2 prong chargers leaking ac current to output via the filter and Y capacitors.
There you go, tested with a genuine Apple 2 prong USB charger.
Result is under 4uA and indeed i can't feel anything touching the ground wire and earth at the same time, the current would have to be at least 50x larger for that. If you plug this into the US mains with its 110V the current would likely be only half this, so under 2uA. Also yes i tried flipping the charger around in case the Y capacitor went to only one prong.
EDIT: But yes, if i dig trough some random chinese wallwarts i can find ones with 250uA of leakage. Tho i still have to touch them very lightly at a sharp point to even feel it (or maybe i am just that insensitive to current)