View Full Version : Laptop: SNR versus external power supply
Hi all,
i have the following problem. If i use my laptop with integrated audio, it works quite good as long as i donot connect my external power supply. If i connect the external power supply the SNR drops at least by 5 dB. I am quite sure that there was a thread discussing such problems some times ago here ;-), which does no more exist :-(
Any idea how to overcome this problem?
73, Daniel
FritzWue
27-12-2005, 23:18
I'm not sure but I think with my DELL the problem might be the charging electronic circuit in the battery pack itself.
I really tried to eliminate noise from two different power supplies with a lot of filtering with no success.
So I now have two battery packs to overcome the problem.
Originally posted by df8uo
If i connect the external power supply the SNR drops at least by 5 dB.
Hi Daniel,
let me suggest the following:
Determine whether the reduction in SNR happens via coupling of unwanted signals in the audio frequency range or via RF interference: Can you hear extra noise when you connect the power supply during a normal audio recording? Does Dream's AM audio output change (check with and without a receiver connected) when you listen to noise or an AM station when you connect the power supply?
Is the SNR already reduced when you connect only the power supply ground to laptop ground? If so, the interference is via the RF path and you should be able to remove it by good filtering (many windings on a high-permeability core in both DC and audio lines).
If the interference enters the laptop via the audio path, then in case of capacitive coupling into the input, a resistor in parallel to the input might help to reduce interference.
To determine the influence of the charging circuit: What happens if you run the laptop without the battery inserted?
Roland
Same problem here. I use a modified Yaesu FP-8 power supply for my notebook, but there are still heavy interferences especially on MW. They dissapear if FP-8 is switched off, but still connected. Also no interferences in battery mode. And there is still interference without the battery pack.
My final Idea to get rid of these interferences is, to connect a power supply directly to the connector of the battery pack. Maby there are some switching regulators that reduce the 19V of the external power supply to the 14.5 V of the battery pack and that causes the interference.
But maybe the battery pack sends an ID to the notebook, so it could afford more than just GND and V+
Originally posted by carknue
Maby there are some switching regulators that reduce the 19V of the external power supply to the 14.5 V of the battery pack and that causes the interference.
But maybe the battery pack sends an ID to the notebook, so it could afford more than just GND and V+
Today's notebook batteries are so-called "smart batteries" which incorporate a charge (micro)processor that keeps track of the state of the battery and which is responsible for charge control via power transistors also contained in the battery pack. There are smart batteries that communicate with the device they are built into via a standardized or a proprietory data bus. In any case, the electronics inside the battery pack includes timing circuits that may generate interference.
I guess that, unfortunately, making the battery RF interference free, is the last thought spent by the engineers designing such batteries...
Roland
At least in my case, the battery pack does not produce these interferences.
The only question is, if the notebook would run with external power on the battery pack connector without bus communication.
Hi all,
tnx for all hints. I will try them out and report any success. However due to the fact that the battery works at least for 3 hours, i think it is the easiest solution for the moment to work on batteries ;-)
73, Daniel
Since the beginning of January when I put a new ACER notebook with genuine "no name" AC Adapter into use I could see much noise on the low end of the Input spectrum and register a SNR lowered by up to 4 dB when AC is connected.
As the laptop is exclusive for DRM decoding (whilst all other work is done on my nice Apple PowerBook :cool: ) I bought a cheap model, the battery of which lasts only for one hour something :(
So tonight I tried another AC adaptor, which seems to be much less noisy, and which was saved from a scrapped Siemens Fujitsu Laptop. Now I can see only a slight decrease of 1 dB if any, when working on AC.
Hello,
This will be about AC adapter problems and other interference sources.
My main computer is also an Acer laptop from 2003, it runs on the dedicated Acer AC power adapter and, depending of the time of the day and the frequency, it ruins reception or decreases SNR of generally 4 to 10 dB. For VOR on 5810 kHz it means a difference between 'no signal seen over the noise' and 20dB SNR (!). For BBCWS on 9410 kHz or DW-Radio on 15440 kHz, it changes almost nothing, signal stays in the 20-24 dB range. Just the stability of the signal is affected.
Of course, with an 'old' Pentium4 processor, with soundcard and LCD panel light switched off, the battery life is only around 90 minutes DRM logging. Battery charge is around 2 hours when the laptop is switched off. That limits my ability to do logs for 3965 kHz and the 49m outlets.
Now, other problems : there are so many QRM sources in the house that I can't also fight them all : fluorescent lights, electricity wires in the walls which interfere a lot with shortwave when just simple lights are on, TV sets, telephone cable that I sometimes forget to unplug from the laptop. That is even worse for me than for shortwave AM listening.
This weekend I was alone in the house, so I experimented DRM listening in various rooms (kitchen,...) : the result is simple, it is better to let the DRM receiver in a QRM-free room, with the laptop totally isolated from all networks or AC power sources around. Then, if that does not ruin the SNR (and that sometimes does !), I plug the soundcard output to my hi-fi system and/or an UHF transmitter, then I can listen to DRM on wireless headphones while cooking, gardening, anything... It is far, far better than letting the laptop in the kitchen or living-room itself.
I really did a good overnight log on 75m-band in the living-room, with fluorescent lights, TVset, VCR and others running. Great. But 6085 or 5990 or 6095 or 6175 or 7295 wouldn't work during the day (compared to QRM-free room). They weren't stable, because of all QRM sources around. Too many dropouts. 9470 and 15440 were good, but really really at the decoding limit, they wouldn't work with the AC power adapter plugged to the laptop...
The other places where I tried before to listen to DRM weren't really successful (mostly hotels in Berlin, Rügen and Harz), but I wasn't yet expeienced in the QRM-fight. With the laptop running on batteries, maybe I would have been able to listen to Putbus 729 kHz when I lived only 15 km away of it. I didn't know it yet. All I know is there was one place I could decode only one second of DRM in one week: my parents' house, close to Paris, in the country but close to High Voltage lines. Except for strong signals, AM was really bad on shortwave there above 15 MHz. But now I know it's even worse on all spectrum with DRM. I don't know how many houses (= potential future DRM consumers) are in such case.
All that to say that I am really worried if the project of DRM promoters is launching new stations on DRM for the average listener who will listen to it while being in the kitchen, in his car... Maybe I am wrong because the reception with my Digital World Traveller and a laptop is less efficient than with the future DAB/DRM sets... I really hope I am wrong, unless this system will only work under traditional DXing conditions, for the actual shortwave listeners. Not for the average consumers who won't fight all interference sources.
Oh, and last but not least, something that appeared to me during this (cold) winter : what will be the average consumer's choice ? 1/ putting his ceramic heater on to warm his house 2/ listening to his new favourite programme on his new DigitalRadio set ? It is really hard for me to do both in the same time, even when there are 2 stairs between the ceramic heater and the shortwave receiver. Do the new DRM receivers' tests include listening to DRM services in an average house with only a telescopic antenna, while doing everyday's things that could cause lots of QRM ? I hope so. Unless there's no need to start new stations for DRM, it's better to save money for the existing programming.
If all I said is because I listen to DRM on a laptop with a simple DWT receiver, and because future DRM sets will be far better, I will be happy I am wrong now. :-)
73, de Stephane, daily and happy DRM listener
Hi Stephane,
you should know that using a simple short wire antenna or any kind of antenna that is unsymmetrical (ie no dipole or loop) will make the current from your wire flow back through the "counterpoise" or electrical ground which, in a house, is made up of all kind of wiring, water tubes, power cables and the like. The ground terminal of the receiver will be connected more or less to this "ground" system, either directly via the case and the protective earth or via some quite large capacitance. Power lines, water tubes, telephone and other wiring effectively become part of your antenna system. With all the interference sources in the house being also connected to this ground system, currents from these interference sources will also flow through this system and will be coupled into your receiver.
So the answer to the problem is to not use an unsymmetrical antenna but to use a symmetrical one instead where the current does not flow back through the ground. Use a loop, a ferrite antenna or a dipole and there will be less interaction with interference sources.
Another problem lies in the fact that because almost no one is listening to AM anymore, interference generated by all kinds of electronic devices has been neglected in the past few years to a larger extent than it used to be when AM broadcasts were still listened to. Also, the number of electronic devices in the average house has increased a lot and switchmode power supplies are very common now for any kind of application, even for powering low-voltage lighting systems (I do also suffer from such an even expensive one which my neighbour owns).
The manufacturers of such devices often look more at the parts count and the additional cost that would be caused by proper filtering and proper EMC (electromagnetic compatibility) performance. As long as no one complains, they won't have to bother. And in almost all cases, no one has complained so far. But, there are still regulations (e.g. EU directives) and standards that will have to be adhered to, otherwise the CE sign that they all put on their devices is just worthless. They might even be fined for putting it there when it comes out that their product exceeds all limits. But someone has to complain at the proper place!
A still much greater danger is coming from other sources: Since powerline communications systems (either in-house or externally fed) cannot be made not to radiate a strong signal which will kill 99% of AM and shortwave signals, the people who still want to make a profit out of those, are actually putting pressure on regulators to increase the allowed levels of interference.
They will tell the politicians (who might even share some of the financial benefit, who knows :confused:) that such interference does no harm, since anyway, no one listens to old-fashioned AM broadcasts in a 70 year old system nowadays anymore and that access to the internet is far more in demand than is the desire to listen to low quality, old-fashioned AM radio. Some politicians in some countries might even find it quite smart, if people are only able to listen to strong local stations which are largely under their control...
But what will happen, if someone who has a strong interference source being operated by one of his neighbours (without the neighbour even knowing), buys a new and expensive DRM receiver and finds out that it just doesn't work? He might believe it to be defective and return it to the shop... But what will happen if a larger number of people do so?
Roland
hello Roland, and all,
You are very true, I cannot expect great reception conditions using the simple wire of my Digital World Traveller. I don't plan yet to build any serious antenna as I have nothing to do it and no experience. And I know how the computer itself is a source of great damage to radio listening. When I compare AM reception on my Sony ICFSW7600G and on the same frequency on my DWT, it is as if I listened to the 10 euros SW-receiver I bought at the 'Plus' supermarket, and often worse. But I'd rather think it's good to try sometimes to catch DRM in such bad conditions, with a bad antenna, as I imagine it reflects how people will commonly use their new DRM receivers.
My wish is really the future DigitalRadio receivers will be as efficient as the Sony one (at least they'll have a ferrite antenna built inside). With still the problem of interferences generated by other electronic devices, power lines, metal-based house architecture... I really fear too the average consumer's reaction to this problem.
73, de Stephane
I have also had a problem QRM from my notepad/power supply/mains. Like many people I have had to resort to powering the notepad off batteries. I have ascertained that the interference is not picked up by the whip antenna I use as it disappears when unplugging the soundcard lead at the receiver. Have tried using 10mH chokes on each connection. This has helped to some extent but not completely cured the problem. I am thinking an optically coupled solution would be best. Does anyone know if it is possible to use an optocoupler to couple an analogue signal to the soundcard?
Richard
I am thinking an optically coupled solution would be best. Does anyone know if it is possible to use an optocoupler to couple an analogue signal to the soundcard?
No, this won't work without modulating the analogue signal onto a pulsed carrier as an FM or a PWM (pulse width modulated) signal before sending it into a light emitting diode. Trying with an analogue signal directly into the diode will lead to a lot of distortion since the diode light output does not depend on input voltage in a linear way. If at all, the analogue signal would have to drive a current source feeding the light emitting diode and even then, the dynamic range should be rather low.
There are special solutions in which a second receiving diode is used to compensate non-linearity but still, there will be a certain capacitance of a few picofarads between input and output through which an RF interference signal may pass to some extent.
I think the problem has to be solved by extensive use of bifilar wound ferrites working as common mode chokes or more efficiently by using a (preferably outdoor) antenna that is decoupled from the ground system carrying all the interference currents.
Roland
Hi Roland, thanks for your suggestions.
Had a feeling that optocouplers were basically intended for data. A PWM link through one is probably worth considering though. Remember I once built one around a 555 timer. Think I just used a simple low pass filter to recover the audio. Have lost the circuit but looks like there may be some on the web. Just hope the noise and distortion caused by the additional circuitry is no greater than that caused by the power supply!
Cheers, Richard
Just hope the noise and distortion caused by the additional circuitry is no greater than that caused by the power supply!
I guess you won't get the dynamic range required.
Also, instead of optocouplers which still have some capacitance between input and output, an infrared fibre transmitter and a fibre receiver connected through a piece of cheap plastic fibre should be used.
Another idea, if one uses the DWT receiver, would be to just send the digital signals of the USB connection via fibres. Since USB is a bidirectional bus, the solution is not quite simple though. Of course the fibre transceivers and their drivers would have to be properly shielded and decoupled in order not to introduce a new source of interference. The DWT would then have to be powered locally instead of through the USB port.
Such USB optical links are also available commercially, eg here (http://opticis.com/product_5.htm).
Roland
introduce a new source of interference. The DWT would then have to be powered locally instead of through the USB port.
Such USB optical links are also available commercially, eg here (http://opticis.com/product_5.htm).
Roland
That looks promising. Where could I buy that? I think that they don't sell to private persons in low quantities.
Even with my new Degen TG33 and Yaesu Fp-8 Power supply, I'm not able to get rid of the interference on MF with DWT and Notebook if not running on batteries only. Hf is not a problem anymore, but MF is still terrible.
@Richard
I use the Soundblaster Audigy 2NX Notebook Soundcard that has an optical Line in and line out. I connected the line out with my old DAT recoder in DA Mode to feed the DRM Audio signal into my hifi system. Before that, I had some audio distortions an humm, which is now completly gone. I'm thinking about trying also the opical input to connect my AOR to the notebook. I see no reason why this should not work. It has got a samplingrate of 48khz. But then there still is the serial connection to the AOR.
I use the Soundblaster Audigy 2NX Notebook Soundcard that has an optical Line in and line out. I connected the line out with my old DAT recoder in DA Mode to feed the DRM Audio signal into my hifi system. Before that, I had some audio distortions an humm, which is now completly gone. I'm thinking about trying also the opical input to connect my AOR to the notebook. I see no reason why this should not work. It has got a samplingrate of 48khz. But then there still is the serial connection to the AOR.
A simple transformer would remove the hum equally well.
The optical connection you are using, is an S/PDIF one, on which an already digitized signal travels and it needs an A/D converter and a digital converter to S/PDIF somewhere ahead of the optical transmitter (and the reverse at the other end).
What Richard was originally looking for was a way to send the analogue signal over an optical link without the need for prior A/D conversion.
But why don't you all build a screened and well-filtered box for your notebook power supplies? A well filtered linear regulator with good RF decoupling of input and output via large common mode chokes will do as well and is much cheaper than the fibre solution which might even introduce some new sources of interference.
Roland
Hi all,
today I found time to compare my Degen Loop against the built in telescopic antenna of my Sangean radio with the laptop powered from the rechargeable batteries or the power supply connected to a laptop (Fujitsu Siemens).
The Degen Loop was mounted at the window, the radio was placed at the same window, the laptop 1,5 m away. Frequency 3995 kHz, time around 17:00
The result is very amazing and confirms what Roland wrote some posts ago.
Sorry for the quality of the attachment but I had to take a photo of the SNR-curve as I could not make a screen shot with this laptop (???!!).
1 = Loop: Sangean and Laptop powered from batteries (SNR limited by the Sangean!)
2 = Loop: Laptop powered from the power supply
3 = telescopic antenna: Laptop powered from batteries
4 = telescopic antenna: Laptop powered from the power supply - the worst!!
5 = 1 (for comparision to exclude the influence of conditions)
6 = Loop: Sangean power from the original AC-adaptor
7 = telescopic antenna: Sangean power from the original AC-adaptor
8 = 1
Powering the Sangean from the AC-adaptor always reduces the SNR independent from the used antenna. Perhaps a noise reducing capacitor at the rectifier is missing. I have to check once I have managed to open the special safety screws preventing to open the housing!!!!
Bernd, DF9RB
6 = Loop: Sangean power from the original AC-adaptor
7 = telescopic antenna: Sangean power from the original AC-adaptor
Hi Bernd,
I would have expected the loop to be better than the telescopic antenna, so that 6 and 7 are reversed.
I guess the reason behind this being that the Degen loop is not really balanced since receiver ground is connected directly to one side of the loop without any further balancing device (coupling loop, balanced-to-unbalanced transformer etc).
There is another effect which has to be taken into account:
Since the telescopic antenna becomes much more effective with regard to received signal voltage, when a "ground" or counterpoise connection is made (capacitively to the mains via the power supply), intermodulation products will also be stronger then. So for a meaningful comparison, the signal at the receiver input would have to be reduced to the same level as in the case when the receiver was battery powered.
Something else:
The Degen loop should be removed when testing the telescopic aerial since it is a resonant circuit tuned to the receiving frequency which might couple into the telescopic aerial.
Roland
Yes have noticed the big increase in signal level together with noise when plugging into the soundcard. After reading all the ideas I think the simplest approach for me is use a balanced amplifier between the receiver and antenna. As Roland says this will help isolate the power supply.
Apart from using a loop maybe another idea is to connect each side of the amplifier to separate halves of a dipole antenna. Presumably if the input impedance of the amplifier is sufficient the dipole need not be big, perhaps just two short lengths of wire. Maybe a matched FET pair or a high frequency op-amp would do the job?
Richard
Apart from using a loop maybe another idea is to connect each side of the amplifier to separate halves of a dipole antenna. Presumably if the input impedance of the amplifier is sufficient the dipole need not be big, perhaps just two short lengths of wire. Maybe a matched FET pair or a high frequency op-amp would do the job?
I suppose it would. Such antennas do exist, often in the form of short dipoles with capacitive end loading using rather large disks at the outer ends which serve to increase current along the conductors. Such an antenna would preferably have to be placed far away from metal structures to prevent distorting the symmetry.
But matching a tuned loop is far more effective. There are other advantages of the tuned loop, namely its selectivity which helps reduce receiver intermodulation. Also the loop, whether broadband or tuned, is mostly sensitive to the magnetic field component, thus the term magnetic loop, which is often used. This makes no difference for a far away signal source, of which the magnetic and electric field components appear in a fixed ratio but it attenuates the influence of local noise which often appears primarily in the form of electric fields. This is because the ratio of electric to magnetic fields near a radiating source is different from fields the source of which is located far (ie more than a few wavelengths) away and it depends on how the field is being radiated.
Roland
Pity the HF loop, in the form of a ferrite rod antenna, is not used more by receiver manufacturers. Presumably the main deficiency is that it tends to produce a lower output signal compared to the whip, requiring an additional rf amplifier stage.
If manufacturers want to implement a whip I wonder if they may be better using a separate ground plate, isolated from the power supply. As people say it is normally impossible to use a cheap receiver connected to the power supply without getting a significant loss in signal to noise ratio. Maybe this would be a possible way for the manufactures to improve sets, although as you say the loop is an inherently better design.
Richard
But why don't you all build a screened and well-filtered box for your notebook power supplies? A well filtered linear regulator with good RF decoupling of input and output via large common mode chokes will do as well and is much cheaper than the fibre solution which might even introduce some new sources of interference.
Roland
I built a PSU inside the screened box my DRT1 was in, then built a separate one inside an empty plug top box with 8 volt regulator, chokes & tantalum capacitors for decoupling. Big improvement in SNR under more difficult conditions. It is possibly like the one which Sat Service have on their web site for the DRT1.
Greetings
I DX - ed low power and distant radiostations only with my
valves Q Multiplyer (preselector).
Q Multiplyer works good for different open air antenas
Regen - Autodyne - DC - Receiver - Q Multiplyer (preselector)
500khz ... 29500 khz with frequency counter
http://www.smail.lt/~ncss/regen_info/My_DRM_system.JPG
http://www.smail.lt/~ncss/regen_info/MySO_Regen_Autodyne_p1.jpg
http://www.smail.lt/~ncss/regen_info/MySO_Regen_Autodyne_p2.jpg
Only with 80m delta loop and Q Multiplyer and ICF7600G i I can good
receive MW Europe DRM radiostations without interefence
in Lithuania at evening
Q Multiplyer can make from FET and low power source from batery.
One duration - Q Multiplyer need adjustment when change
receiving frequency
ICF7600G front - end sensitivity and selectivity grow up many times
when i use it with Q Multiplyer (preselector) and noise decrease.
Best regards;
Saulius
If manufacturers want to implement a whip I wonder if they may be better using a separate ground plate, isolated from the power supply.
Hi Richard,
to be effective, the separate ground plate would have to be bigger or to be precise, to exhibit a lower ground impedance than the virtual ground which is introduced when a capacitive connection to the mains wiring is made. So this means that a metal plate with a size of several wavelengths or a number of tuned quarter-wave long radials would be required! Quite impractical! If you happen to have a large metal covered roof, put a vertical stick with a high input impedance amplifier in the centre of it and connect the shield of your coaxial cable to the metal plate. This would be perfect!
Since there are only few people who can do that, a small loop is a much smarter solution but the problem is more a psychological one since ordinary radio users have become used to whip antennas for the last decades rather than to loops.
BTW: In a car with no larger metal wiring structures like mains wiring to connect to however, a vertical stick mounted over the metal structure works very well.
Roland
Roland, maybe we have different ideas over what constitutes a ground plain. In a typical receiver I would consider the ground plain to be the receiver chassis(power supply rails), this is like the metal roof. The telescopic antenna is a smaller version of the vertical you describe. The difference in electric field between these two parts obviously creates the electrical signal.
This is the arrangement which works when the receiver is run off batteries. When connected to the mains there is going to be a lower impedance path to ground, as you describe, and a corresponding increase in signal - the size of the ground plain has been increased.
However my receiver still works well and picks up less noise when not coupled to true ground. My conclusion is that a separate ground plain would only need to be similar in size to the receiver chassis inorder to give the same performance from a short telescopic antenna.
Richard
Hi Richard,
Roland, maybe we have different ideas over what constitutes a ground plane. In a typical receiver I would consider the ground plane to be the receiver chassis (power supply rails), this is like the metal roof. The telescopic antenna is a smaller version of the vertical you describe. The difference in electric field between these two parts obviously creates the electrical signal.
This is the arrangement which works when the receiver is run off batteries.
so far, I agree.
When connected to the mains there is going to be a lower impedance path to ground, as you describe, and a corresponding increase in signal - the size of the ground plane has been increased.
Exactly. However, since the connection is not to ideal ground, there is a certain impedance to ideal ground and there are also interference voltage and perhaps also current generators of all kind you will find in a typical house between ideal ground and receiver chassis.
These are ones we want to avoid. There are several ways to do that:
at the source (not always within one's reach) by preventing them to produce the interference signal
by isolating the interference sources from ground at the source
by isolating the receiver from virtual (mains) ground and simultaneously
by providing a lower impedance path to ideal ground (of lower impedance than via the mains) which will short the interference voltage
My conclusion is that a separate ground plane would only need to be similar in size to the receiver chassis in order to give the same performance from a short telescopic antenna.
If the impedance of the mains wiring to ideal ground is lower than that of your metal plate (most likely, since your comparably small metal plate will more or less only constitute a capacitor between the receiver ground and the mains wiring that is around it), then the current will pass via the mains wiring and this assumption won't be true.
Roland
Roland, I am guessing this antenna arrangement will be less noisily, although as you say the critical point is the effect that capacitive coupling between the ground plate/telescopic rod and the noisy parts in the receiver would have. It is difficult to predict this without experimentation. Will see if just positioning the soundcard lead close to my telescopic battery powered receiver is enough to cause a drop in SNR.
With a dipole/loop antenna arrangement connected to a balance line a solution would be to position it away from the receiver, and any other sources of noise of course!
Richard
Hi,
I have looked around a bit on the net and found this about interference from computers and power supplies and how to reduce it:
http://www.arrl.org/tis/info/rficomp.html
and the link from there to
http://www.arrl.org/tis/info/cable-modems.html
Note that, if in Europe, you don't have to buy expensive US-made ferrite cores, similar high-permeability materials like Epcos (Ex-Siemens) N20, N27, N30 materials or similar will do as well. If you can't remove the connector in order to stick the cable through the core, use a pair of E cores, wind your cable on it and then close the core.
There's a bit of generalized information here:
http://www.radioing.com/eengineer/intro.html
Also check out this page and a few pdfs on the topic that can be found there:
http://www.elmac.co.uk/papers01.htm
Roland
vBulletin v3.5.3, Copyright ©2000-2010, Jelsoft Enterprises Ltd.