Ask the Internet: DSL Service Question

The problem with cable internet choking at peak times is mostly a thing of the past - that was an issue in the early days, but now the networks are segmented much more efficiently so that doesn't happen anymore to any noticeable degree. And with the 15-20 Mbps speed of today's cable, you wouldn't likely notice the bottleneck even if there was one.

As for the DSL modem "burning out"...um...no. I'm pretty sure the guy who told you that was drunk.

Apologies for not really being able to answer your question, but I'll offer my experiences in case they're helpful. The whole thing smells like a crock to me. "Built like a tank" sounds at best like it's from the category of "lies we tell to children". In any case the answer would be completely different now since the current modems will be different than what was available a couple years ago.

We had DSL for about ten years, up until last year. We never got more than 768kbps. AT&T rolled fiber (U-Verse) out to within a block of my house but stalled there two years ago. I finally gave up waiting for it and switched to cable last fall. I'm seeing real-world downloads around 2Mbps most times, and haven't seen any neighborhood congestion issues. In general it's been more reliable than the DSL; had a couple strange DNS-based failures but they clear up within 30m or so. YMMV of course. Interestingly, the cable modem (which also handles home phone) seems to have a battery backup; not sure how long it would last, however.

So I've been happy with the cable internet service. Don't get me started on switching from Tivo to the cable DVR, though. Ugh.

He said that the older modem can handle it because it's "built like a tank," but that a newer one would burn out very quickly at that speed.

Yeah, more bits per second causes a lot of friction heat and eventually spontaneous combustion, if the thing is poorly made.

(Um, no. No it won't.)

And the cable experience is probably going to depend very much on the neighborhood and the cable company you've got. I used to experience a slowdown in the evening hours, but had never noticed it until I stayed home sick for a few days. Which is to say, it was more than adequate in the evenings but blisteringly fast during the day. Today, I don't see any noticeable slowdown, much as Jeff, above me, suggests.

It's still a gamble, but frankly so is DSL. Today they're both relatively good gambles, I think.

Um, what, burn out? Are DSL modems in the US coal powered or something? He is a hire from the local power company, and he has a problem switching mental gears from power transmission to data?

It sounds like complete bunk to me.

DSL modems have a limited supply of electrons. The faster your connection speed, the faster the reservoir of electrons in the modem is drained. When it empties, the modem's electron pump tends to overheat. In any well-built modem, the empty reservoir / overheating pump triggers a shutdown.

The sensors needed to detect empty reservoirs / overheating pumps are expensive. So modern modems simply use a fuse. The fuse will eventually, for any of a number of reasons, blow. The modem effectively shuts down. This normally works, but can go wrong. It's not unknown for the modem to catch fire before the fuse blows.

This is why it's important to use older sturdier modemsâor more accurately, modems that use sensors instead of fusesâand to recognize claptrap when the salesman tries to push a new fused modem on you.

This comment may be complete horseshite. Any use of the information presented is not only at your own risk, but will be to the considerable amusement of others. Results will vary, depending on the size of your gullibility. Contains nuts.

Higher speeds generally mean more heat--at least that's true of internal clocks in computers, so it's plausible that there would be a similar effect in modems. But if 1.1Mbps vs. 1.5Mbps is enough to make the difference between viable service and your modem frying itself, that doesn't speak well of the engineers who designed that modem. Unfortunately, that is a valid concern--hard as it may seem to believe, Janne, DSL modems do burn out. I had a DSL modem die after about five years of service (originally Verizon, who sold their lines in northern New England to FairPoint), and my mother also observed this symptom (also Verizon; she reacted by switching to cable).

As others have said, your best option depends on who your cable company is. If it's Comcast, run (don't walk) to any other option--they are my local cable provider, which is one reason (along with my lack of a TV set) I have never considered buying their internet service, and they have a reputation for being evil even by the standards of telecom companies. I don't know anything about Time Warner's track record.

At higher speed and greater distance the line driving amplifier in the modem does have to "work harder" (larger signal amplitude into greater capacitance), which means it will run hotter. A poor design with inadequate attention to heat dissipation could result in premature death of the amplifier. Because DSL is asymmetric (ADSL), the data rate from the phone office to you is greater than that from you to the office. This means that the line driver that has to do the hardest work is the one in the phone office, not the one in your modem. Usually, your transmission rate will be some more or less fixed fraction of your receiving rate, so higher nominal rate will mean somewhat higher power dissipation in your modem too. The line is constantly driven, so even if you aren't sending "real" data, the line driver is still hard at work.

I have ADSL running at nominally FIFTEEN megabits per second receiving speed and about a tenth of that for sending speed. I have a five year old 2Wire modem and I'm about 5-6 kilometers from the phone office. The modem is the same one I used for my original 1.5Mbps service.

It sounds like bunk to me, but it is true that the big ISPs usually issue you with the cheapest, crappiest DSL modem that they can obtain in bulk (Google for reviews of the model number to check). Some ISPs will also charge you a monthly "rental" fee for the modem that they have supplied, so you might want to check your bill to see if this is the case. If it is the case, and if your problem really is the modem, then you might be better served by buying your own DSL modem ($50 should suffice) from somewhere like Office Depot or Staples. Make sure you can take it back the next day if it does not improve your connection.

P.S. If you are paying your ISP to provide you with a wireless router as well then you are definitely doing it wrong. You will save money in the long run by buying it independently. You can get a combined modem/router, but I recommend getting them separately so that you can reuse the router if you switch to cable.

I have had DSL (Verizon) for about 5 years. After a few problems getting started, it was fine until about 6 months ago. Couldn't connect to internet sometimes or slow speeds sometimes, especially at high traffic times. We also have a long connection, 13,000 feet, paying for a 3 Mbps and getting 1.5 Mbps, so we get a free wireless modem. Last fall we had multiple visits and a diagnosis of an old, electrically leaky/noisy buried cable (the last 300 yards), so cable was replaced. Didn't help. The last service tech, on a Sunday morning (!) looked at the modem, said it was old (I think it wasn't updatable to the current carrier system). So, without writing one more piece of paper or making another service call, he pulled a new modem off his truck (rather than having me call service again and request a replacement) and swapped the 5 year old modem for a current modem. Works fine now.

Higher speeds generally mean more heat--at least that's true of internal clocks in computers, so it's plausible that there would be a similar effect in modems. But if 1.1Mbps vs. 1.5Mbps is enough to make the difference between viable service and your modem frying itself, that doesn't speak well of the engineers who designed that modem.

We're talking about the modem provided by Verizon, so there's a decent chance it was designed and built by monkeys. poorly-trained monkeys.

The speed difference might be 11 vs 15, rather than 1.1 vs. 1.5-- I don't recall the exact order of magnitude, and can't exactly check it with the modem being dead...

At higher speed and greater distance the line driving amplifier in the modem does have to "work harder" (larger signal amplitude into greater capacitance), which means it will run hotter. A poor design with inadequate attention to heat dissipation could result in premature death of the amplifier. Because DSL is asymmetric (ADSL), the data rate from the phone office to you is greater than that from you to the office. This means that the line driver that has to do the hardest work is the one in the phone office, not the one in your modem. Usually, your transmission rate will be some more or less fixed fraction of your receiving rate, so higher nominal rate will mean somewhat higher power dissipation in your modem too. The line is constantly driven, so even if you aren't sending "real" data, the line driver is still hard at work.

This sounds like the more detailed explanation I was given (and then mostly forgot) when I expressed skepticism about the burnout thing a couple of years ago.

Speaking as someone who designs RF electronics-- and worries obsessively about thermal management in environments far more malicious than your desk or even your packed, 110 degree Fahrenheit utility closet-- let me make some serious observations about thermal design issues.

1) Yes, it is possible in some circumstances (bit rate, low incoming signal, etc) that a modem might generate more heat than in other circumstances. Theoretically.

2) However, in ordinary, day to day use, those situations should not be a big deal. 1.1 MBPS vs 1.4 MBPS? Pfah. A low noise amp doing a little extra work? Pfah.

3) And in any case, there are ratings and testing to pass. If the thing is sold as a 1.5 MBPS device, it ought to effing work at 1.5 MBPS. Period.

4) Let's also distinguish between, "It broke," and "It burned out due to thermal mismanagement." On the one hand, these effects are real, and they are a great pitfall for stupid designers to fall into: Witness the older X-Boxes and their Red Rings of Doom. Those are pretty conclusively exposed as thermal mismanagement problems, brought about by a software company stubbornly insisting that hardware design was easy. On the other hand, without doing an autopsy on a dead device, it's not easy to tell "It broke," from "Bad thermal management." In no case do I trust Verne The Cable Guy to make that distinction for me.

5) If your modem is not a cheap piece of crap, and you're not storing it in your (lit) fireplace, thermal management ought not to be an issue.

We didn't have problems with our modem burning out with DSL, but we did have all sorts of problems with our DSL being flaking for various reasons, and speed was an issue because of distance from the central office. We had problems ranging from squirrels chewing on wires in a box somewhere between us and the CO to having to move the modem because if it was too close to my husband's CRT, every time his CRT woke up, it took down the network briefly.

So, when we moved, we decided to give cable a try, and have had zero problems in the 7+ years since. Even during the 4-day power outage we had a few years ago, our internet stayed up (we used a UPS sparingly to grab stuff). I've not noticed any slowdown in the evenings.

You missed a real opportunity there. When he said THEIR newer modem would burn out, you should have thanked him for telling you in advance that their equipment was junk and that you should switch to a better vendor.

That would get the supervisor's attention.

But I can't believe you (or your dog) didn't engage him in a quantum physics discussion about the cause for modem burnout.

You missed a real opportunity there. When he said THEIR newer modem would burn out, you should have thanked him for telling you in advance that their equipment was junk and that you should switch to a better vendor.

That would get the supervisor's attention.

This was a conversation with the tech who came to the house to fix our previous problem. Since it had taken three tries to get anybody to the house in the first place, I wasn't going to say anything that might cause him to stomp off in a huff, because God only knows when we'd get another service call...

I didn't pay too much attention to it, either, because he had managed to fix the problem that led to the call. (The previous owners had put a low-pass filter on the phone line, which was blocking the frequencies used for DSL. We hadn't noticed previously because the room we used to use as an office had had a separate phone jack run to it, that was split off before the filter. He removed the filter, and everything worked great after that...)

DSL actually works by beetles running up and down the line carrying packets. This is why it takes so long to get started after you have paid -- the beetles have to chew their way to you from the head office, homing in on the modem. But once they reach you there's a continuous chain of DSL beetles scurrying along with packets, a bit like leafcutter ants.

This isn't strictly technically true, but it's a useful visualisation that covers the usual failure modes -- slow beetles, distracted beetles, beetles piling up. It also explains why you have to switch the modem on and off sometimes to restore the homing signal.

This is friendlier than the visualisation of the Demon Internet Gateway I used to dial-up to with some trepidation when I lived in the UK.

Did run into a heating problem in mine once that would cause a reset and at times a shutdown. Curiously the problem eventually manifested itself by the power outlet supplying it failing. Once the outlet was replaced, the modem worked without problem. Insufficient current through the resistive outlet.

Higher speeds, all things being equal, creates more heat. Many earlier modems used large heat sinks to handle the heat. These units were often larger, usually noticeably heavier and they handled thermal abuse better. Not uncommon to hear them talked about as 'built like a tank' or about how you could 'hammer nails with them'.

Later modems, using higher efficiency chip-sets that produce a little less heat have shifted to smaller and lighter, or no, heat sinks. This makes the modems cheaper and easier to produce and cheaper to ship.

A lot depends on the installation. If the modem is in a cold location and well ventilated the smaller heat sinks might do fine. If the unit is in a location that gets hot or where there is little ventilation it will fail earlier, become less reliable, or, by way of internal throttling tied to chip temperature, slow down. Often some combination of all three.

Theoretically, if the newer models were designed to be cheaper and they skimped on adequate cooling for the system, damage could be don ein the long term.

Firmware update sometimes breathe new life into the modem.