Kitz Forum
Broadband Related => Broadband Hardware => Topic started by: digitalnemesis on October 17, 2020, 08:24:27 PM
-
What are the consequences of using a 12V 1.5A instead of a 12V 2A adapter on a router?
-
If the router ever draws more than 18 watts bad things may happen.
-
Certainly if using a Zyxel router bad things WILL happen, those with 5ghz wifi need 2A. I speak from experience.
Stuart
-
Personally I would guesstimate that you can lose say 200mA per USB 2.0 port not occupied, 400mA per USB 3.0 port.
I've certainly managed to use a TP-Link WDR3600 on only 1A PoE adapter when it came with a 1.5A PSU which I accounted to the fact it has two USB 2.0 ports that need power if used. For reference when sat idle acting as a switch only (theoretically WiFi is on but I'm not seeing the SSIDs so I think OpenWRT is being funny and its not broadcasting) its pulling only 5.3W using the PoE adapter I have it on, a dummy CPU load pushed it to 6W.
There are really no guarantees though and using an under-rated PSU potentially if its a badly made PSU you could fry it if you overload it.
-
The most likely scenario is that as the load drawn from the PSU increases, the output voltage will sag from the specified 12V DC.
Dependant upon how deep is the sag, you may well find that the modem/router will re-boot. And then continue in a perpetual re-boot cycle. That will definitely make the DLM process unhappy.
-
What Burakkucat said; the modem-router will get more and more unreliable and will crash more and more often the lower the voltage droops. The voltage droops could be of very short duration.
-
I'd personally risk in bridge/modem mode (its unlikely to go over 500mA), but not in full router mode where the load will be much more higher.
-
I’m not following Alex’s reasoning, :) but then I’ve never looked into such a thing so what do I know. I do recall that debugging such problems, especially short spikes is awkward and fixing them cures a lot of bizarre faults.
-
I’m not following Alex’s reasoning, :) but then I’ve never looked into such a thing so what do I know.
Manufacturer PSUs should always be rated higher than the maximum current the device will ever need. If you're only using a subset of the hardware, it will NEVER hit that current requirement.
eg If it has two USB ports you'll never use, thats 5W that will never be drawn from the PSU. Not using the WiFi might save another 5W.
Is it recommended, no, but there are situations (my own where PoE to 12V adapters outputting 1A are common and 2A are much more rare/expensive) where its a worthwhile compromise.
That said, if you're talking about powering from the mains, just get the right PSU, they are neither rare nor expensive. But as a temporary stop-gap while waiting for a replacement, again it doable.
-
There is also the fact that some PSUs are tightly regulated, and some are not.
A regulated Power brick rated at 12V 1A can be expected to provide 12V, at up to 1A. It’ll probably work at a bit more than that but will be over-stressed and likely fail prematurely
An unregulated brick with similar rating is expected to produce 12V when the current drawn is exactly 1A, with different voltages at different currents. Should the current drawn exceed rated current, the voltage would drop below rated voltage. Probably worse, should you draw less current the voltage may rise, with a likelihood of damage to connected equipment.
Personally, I always do my absolute best to obtain original replacement parts when Power bricks fail. Just too many variables. And in the nightmare (but not inconceivable) scenario where one such variable leads to catastrophic failure involving smoke and flames, home insurers may fall back on the fact that the user manual probably warned against using anything other than original parts. :(
-
Its pretty rare to find an unregulated PSU these days as all devices are variable load so tend to use switching PSUs that can more efficiently deliver this capability. But yeah, definitely not a good idea to use a random PSU if you do not know if its regulated or not, although if its from another network appliance it will be.
-
Its pretty rare to find an unregulated PSU these days
I disagree, unregulated PSU bricks are quite common. Random example... the PSU brick for my Openreach-branded ECI FTTC Modem rates itself as 12V 1A. I have just measured the open circuit voltage, it is ~15V. Let's hope I don't get punished by DLM for conducting that little experiment. :D
The exception is probably 5V USB bricks, which need to conform to a fairly tight voltage range in order to call themselves "USB".
Does it matter? Well that depends on the device to which it is attached, and whether the designers were briefed to make it work with unregulated PSUs, or whether they were assured of a regulated power supply. That may be sometimes be documented in device specifications if you are lucky - but don't recall if I have ever seen it on consumer grade documents.
Edit, PS. Interesting discussion in following link as to whether “switch mode” equates to (closely) regulated. Hint... no, it does not.
https://electronics.stackexchange.com/questions/57654/switch-mode-power-supply-and-output-voltage-range-unregulated
-
I disagree, unregulated PSU bricks are quite common. Random example... the PSU brick for my Openreach-branded ECI FTTC Modem rates itself as 12V 1A. I have just measured the open circuit voltage, it is ~15V.
I do not believe that's "technically" unregulated, as switching PSUs by their nature are regulated (and plain unregulated PSUs are big heavy bricks due to needing hefty transformers), just how close they stick to their target can vary a little. However they are designed for a minimum amount of load, so testing unloaded is pointless.
I put mine on a self powered LCD meter and it came out at 14v, so with even that tiny load its bringing it down into range.
-
I didn’t know about the preponderance of unregulated PSUs. Good to know. Rather frightening.
-
@Alex, You seem to be making the mistake of equating “linear == unregulated”, and equating “switch mode == regulated”.
A traditional supply with a big heavy mains transformer, can be very closely regulated - just stick a voltage regulator on the output, or even just a zener diode. Its voltage will be regulated, just not likely to be very efficient.
A switching supply is likely to be much more efficient, but there is nothing in the definition of “switch mode” to mandate that it is closely regulated.
There is nothing “pointless” about testing O/C voltages. Lots of devices consume only tiny currents, nano Amps if not micro Amps when idle or in “sleep mode” yet their supply decoupling capacitors will still be charged to the full voltage... Not a problem if they are designed for it, but a problem if they are not.
Moreover, my “open circuit” measurement was of course a little white lie as my DVM itself drew a small current so not really open circuit. I think input impedance is 10Mohm, iirc.
-
. . . my “open circuit” measurement was of course a little white lie as my DVM itself drew a small current so not really open circuit.
Isn't there an oscilloscope in the seventh layer grotto?
-
Isn't there an oscilloscope in the seventh layer grotto?
There is indeed...
https://forum.kitz.co.uk/index.php/topic,24212.msg407595.html#msg407595
Is there some experiment you would like me to conduct?
-
I was thinking about your
statement confession that the DMM presented a (tiny) load to the switching PSU and any possible way the off-load voltage could be measured . . .
I expect that the input stage of an oscilloscope would posess a greater impedance than any DMM.
-
I expect that the input stage of an oscilloscope would posess a greater impedance than any DMM.
With a decent quality DMM such as Fluke don't think there would be that much difference in input impedance certainly by comparison to an older CRO, development of components such as IG FET's changed the goal posts.
I still own an old analogue AVO 8 Mk.4, 20K.Ohms/Volt input impedance, perfectly fine in it's time for trouble shooting thermionic valve equipment and still gives a better indication of low frequency oscillations than a DMM in lieu of a scope.
Also handy for the old "wet finger" test on transistors.
Apologies, I digress !
-
With a decent quality DMM such as Fluke don't think there would be that much difference in input impedance certainly by comparison to an older CRO, development of components such as IG FET's changed the goal posts.
Thank you. :)
-
My current favourite DVM was purchased from Maplin mid noughties, probably paid £10-£20. I have lost the formal specification sheet but have established, by experimentation with resistor networks, that the impedance appears to be 10M ohms - on the low voltage DC ranges at least.
My little (toy) oscilloscope discussed in linked thread above is documented to have an impedance of 1M ohm.
Observed behaviour is consistent with above. A recent and ongoing project involves detection of very small currents from an LED that is deployed “inside out” as a photodiode. By loading the LED with a high value resistor, the Maplin DVM can be used to measure the photovoltaic current, even under indoor lighting conditions. Providing I make allowance for the 10M ohm load of the DVM itself, by measuring the voltage on the meter’s lowest scale, I can calculate the actual photovoltaic current. The little oscilloscope is pretty useless in comparison - presenting simply too high a load to allow meaningful measurements to be obtained.