No extra processing would have been required. ADCs and DACs at the time already supported a-law and u-law encoding for telephone systems which essentially allow 12+ bits of dynamic range to be encoded into 8 bits.
Now could an accurate higher resolution converter with a logarithmic transfer curve could have been produced? I doubt it and if it could, so what? It would only have saved about 4 bits at the cost of not being able to use existing linear converters.
I have. The half life for CFL bulbs where I am, outside of St. Louis, is about 3 months.
Now this may very well be a poor power quality issue but that is something that is beyond my control short of installing an online UPS or power conditioner for lighting and I already have 5 online UPSes or power conditioners for various other things.
I have run across embedded hardware before where the ethernet ports were *not* assigned MAC addresses and it was the responsibility of the OS to set the MAC address before the ports could be used.
because evaporative cooling (sweating) does not work in space
Evaporative cooling works fine in space and the new NASA spacesuit uses it. It does not work inside of a pressurized spacesuit unless air can be circulated and dried. Skin-tight spacesuits take advantage of it directly to allow the body to regulate its own temperature through evaporation through the skin.
Just to expand on this point, each character in Atlas Shrugged illustrates a specific flaw, especially Dagny Taggart and Hank Rearden, or virtue. Perhaps a better writer could have included more characterization to add to that but I have difficulty finding fault with keeping it simple in the interest of preventing confusion.
Usually I find that the people who do not like the book also missed the point. Dagny and Rearden were the main antagonists.
With enough of them, sure. The fixed power line losses from the generators to the homes are compensated for in one direction. The same problem would occur if enough houses in one area do not use power instead of adding it into the grid.
The LEDs themselves under ideal conditions may last 50k hours but the whole bulb and especially the electronic ballast is subject to complete failure in a much shorter time frame. Where I live, they last a couple months before power line irregularities burn them and CFL bulbs out.
You could have found articles discussing their product in a modicum of time that are prominently linked on the first page of their web site and gotten your answer undiluted.
Summary:
It implements stream based flow control while evaluating the behavior of each stream and penalizing the misbehaving ones.
Traffic shaping on the incoming side is still effective though even given that it has to drop packets that have already been sent over the most expensive part of the link. Dropping packets is the one sure fire way to signal to the transmitter that it should stop sending so quickly and while the server can ignore ECN, it cannot ignore dropped packets.
If the incoming aggregate flow rate is kept below the level of the slowest link which is almost always the customer's link, then the intervening buffers will tend to be depleted minimizing latency. At the very least, the transmit buffer immediately on the other side of the customer's link can be kept at a minimum level.
While it is true that receiving the data again across a slow link is inefficient, dropping packets is the only universal way to signal IP congestion. (ECN) Explicit Congestion Notification can signal congestion at the IP level without dropping packets but of course few devices implement it or perhaps even go out of their way to ignore it in the quest for individual performance at the cost of degrading the network for all other devices.
Traffic shaping does indeed work for outgoing *and* incoming data. Obviously on the incoming side the packets get dropped after the expense of sending them over the slow link but using it does have the advantage of minimizing queue depth and lowering latency.
Absolutely. Optocouplers are slow because of transistor storage time and not the LED which is why if you have access to the base of the transistor, you can make the optocoupler much faster.
I have read stories about mysterious businesses presumably associated with the cable or DSL operators buying tower space to put up WiFi gear with the intention of blocking local WISP operators.
If you look at the FBI Crime Reports, you will see that there are 37 criminal firearm based homicides [fbi.gov] for every self-defense homicide by a civilian [fbi.gov].
So self defense only occurs if the assailant is killed. If they are wounded or simply flee when their intended victim produces a firearm and resists, then that does not count.
I had much the same experience in the 90s. The QIC drives would verify the tape as good but when it was removed, it became unreadable. I switched to a combination of MO and CD.
If you don't have resistance, any capacitive or inductive effects will be 100% efficient, no?
No. Dielectric polarization losses in capacitors and hysteresis and eddy current losses in inductors will still exist when superconductors are used. In high voltage high voltage power transmission lines, corona discharge losses will still exist.
Oscilloscopes make very handy back end modulation analyzers when combined with a demodulator and would also be used in designing the demodulator itself. The common RF applications I see them used for are broadband envelope measurement and broadband RMS measurement where they can often be used to calibrate other instruments.
If you are buying turnkey solutions, then obviously an oscilloscope is of less use since even if you used it to diagnose a problem, you will be reliant on the vendor to fix it. Not every solution is best solved buy outsourcing.
The problem here is that I don't *want* bells and whistles. They are just more things to break. One of the major reasons I prefer small RWD pickups is because they are easier and less expensive to maintain.
It is common for USB to RS-232 adapters (and anything else which runs on low voltages) to produce RS-232 compatible "polarity" and formatting without generating RS-232 voltage levels which will cause problems with real RS-232 receivers. Even worse, since the polarity is compatible, you cannot use a standard logic level to RS-232 level shifter because they are inverting. If you want to do that, then you have to invert the signal yourself although that is not difficult.
Manufacturers like Garmin are very careful to word their instruction manuals and specifications. Their devices do not produce RS-232.
What kind of features would you recommend I look for in a used oscilloscope?
How "used" do you want to go? When buying old oscilloscopes, I would recommend getting two so you can use one to repair the other and you can learn a lot maintaining old equipment like that.
If you get a digital storage oscilloscope, I would not get one that does not support peak detection which rules out some current inexpensive Rigol models. Long record lengths can sort of make up for that lack but never completely. I would rather have a short record length and peak detection instead of a long record length without it.
For signal integrity analysis the sampling rate is almost irrelevant and bandwidth is everything. That is why I lament the lack of available sampling oscilloscopes.
My fastest oscilloscope is 14 GHz but only has a 50 kHz sampling rate and works great for this type of application.
If you are dealing with well characterized high speed digital signal, than presumably you know to use a low impedance passive probe or an active probe. Hopefully you know to minimize ground lead length as well.
Without proper probing technique, a faster oscilloscope is not going to be helpful.
Slashdot Mirror
No extra processing would have been required. ADCs and DACs at the time already supported a-law and u-law encoding for telephone systems which essentially allow 12+ bits of dynamic range to be encoded into 8 bits.
Now could an accurate higher resolution converter with a logarithmic transfer curve could have been produced? I doubt it and if it could, so what? It would only have saved about 4 bits at the cost of not being able to use existing linear converters.
I have. The half life for CFL bulbs where I am, outside of St. Louis, is about 3 months.
Now this may very well be a poor power quality issue but that is something that is beyond my control short of installing an online UPS or power conditioner for lighting and I already have 5 online UPSes or power conditioners for various other things.
Not when they fail more quickly than the incandescent bulbs they replaced.
I have run across embedded hardware before where the ethernet ports were *not* assigned MAC addresses and it was the responsibility of the OS to set the MAC address before the ports could be used.
Just to expand on this point, each character in Atlas Shrugged illustrates a specific flaw, especially Dagny Taggart and Hank Rearden, or virtue. Perhaps a better writer could have included more characterization to add to that but I have difficulty finding fault with keeping it simple in the interest of preventing confusion.
Usually I find that the people who do not like the book also missed the point. Dagny and Rearden were the main antagonists.
PCs and other devices used to include a hardware jumper which had to be switched to allow updating the Flash memory.
With enough of them, sure. The fixed power line losses from the generators to the homes are compensated for in one direction. The same problem would occur if enough houses in one area do not use power instead of adding it into the grid.
The LEDs themselves under ideal conditions may last 50k hours but the whole bulb and especially the electronic ballast is subject to complete failure in a much shorter time frame. Where I live, they last a couple months before power line irregularities burn them and CFL bulbs out.
You could have found articles discussing their product in a modicum of time that are prominently linked on the first page of their web site and gotten your answer undiluted.
Summary:
It implements stream based flow control while evaluating the behavior of each stream and penalizing the misbehaving ones.
Traffic shaping on the incoming side is still effective though even given that it has to drop packets that have already been sent over the most expensive part of the link. Dropping packets is the one sure fire way to signal to the transmitter that it should stop sending so quickly and while the server can ignore ECN, it cannot ignore dropped packets.
If the incoming aggregate flow rate is kept below the level of the slowest link which is almost always the customer's link, then the intervening buffers will tend to be depleted minimizing latency. At the very least, the transmit buffer immediately on the other side of the customer's link can be kept at a minimum level.
While it is true that receiving the data again across a slow link is inefficient, dropping packets is the only universal way to signal IP congestion. (ECN) Explicit Congestion Notification can signal congestion at the IP level without dropping packets but of course few devices implement it or perhaps even go out of their way to ignore it in the quest for individual performance at the cost of degrading the network for all other devices.
http://en.wikipedia.org/wiki/Explicit_Congestion_Notification
Traffic shaping does indeed work for outgoing *and* incoming data. Obviously on the incoming side the packets get dropped after the expense of sending them over the slow link but using it does have the advantage of minimizing queue depth and lowering latency.
Absolutely. Optocouplers are slow because of transistor storage time and not the LED which is why if you have access to the base of the transistor, you can make the optocoupler much faster.
I have read stories about mysterious businesses presumably associated with the cable or DSL operators buying tower space to put up WiFi gear with the intention of blocking local WISP operators.
So self defense only occurs if the assailant is killed. If they are wounded or simply flee when their intended victim produces a firearm and resists, then that does not count.
http://en.wikipedia.org/wiki/Battle_of_Athens_(1946)
You just have to shield the important parts and add common mode chokes and maybe feedthrough capacitors.
I had much the same experience in the 90s. The QIC drives would verify the tape as good but when it was removed, it became unreadable. I switched to a combination of MO and CD.
No. Dielectric polarization losses in capacitors and hysteresis and eddy current losses in inductors will still exist when superconductors are used. In high voltage high voltage power transmission lines, corona discharge losses will still exist.
Oscilloscopes make very handy back end modulation analyzers when combined with a demodulator and would also be used in designing the demodulator itself. The common RF applications I see them used for are broadband envelope measurement and broadband RMS measurement where they can often be used to calibrate other instruments.
If you are buying turnkey solutions, then obviously an oscilloscope is of less use since even if you used it to diagnose a problem, you will be reliant on the vendor to fix it. Not every solution is best solved buy outsourcing.
The problem here is that I don't *want* bells and whistles. They are just more things to break. One of the major reasons I prefer small RWD pickups is because they are easier and less expensive to maintain.
It is common for USB to RS-232 adapters (and anything else which runs on low voltages) to produce RS-232 compatible "polarity" and formatting without generating RS-232 voltage levels which will cause problems with real RS-232 receivers. Even worse, since the polarity is compatible, you cannot use a standard logic level to RS-232 level shifter because they are inverting. If you want to do that, then you have to invert the signal yourself although that is not difficult.
Manufacturers like Garmin are very careful to word their instruction manuals and specifications. Their devices do not produce RS-232.
How "used" do you want to go? When buying old oscilloscopes, I would recommend getting two so you can use one to repair the other and you can learn a lot maintaining old equipment like that.
If you get a digital storage oscilloscope, I would not get one that does not support peak detection which rules out some current inexpensive Rigol models. Long record lengths can sort of make up for that lack but never completely. I would rather have a short record length and peak detection instead of a long record length without it.
For signal integrity analysis the sampling rate is almost irrelevant and bandwidth is everything. That is why I lament the lack of available sampling oscilloscopes.
My fastest oscilloscope is 14 GHz but only has a 50 kHz sampling rate and works great for this type of application.
If you are dealing with well characterized high speed digital signal, than presumably you know to use a low impedance passive probe or an active probe. Hopefully you know to minimize ground lead length as well.
Without proper probing technique, a faster oscilloscope is not going to be helpful.