R
ric
Franc said:Sorry, but none of this makes any sense. See my other post in this
thread.
Not unusual for half_wit...er...half_pint.
Franc said:Sorry, but none of this makes any sense. See my other post in this
thread.
ric said:Not unusual for half_wit...er...half_pint.
ric said:Not unusual for half_wit...er...half_pint.
He means what he said (not too clear).
The OP said he was getting ~38 actually, however
I think you need to re-read my posts where a filesize in bytes is
simply multiplied by 8 to get bits them divided my time to get
a bits per second rate, clearly this will not take into accound start
and stop bits.
I don't think so, as I am pretty sure 99% of people connect at full speed
and any apparent delays in speed are due to there reasons other than the
telpphone line quality
Franc Zabkar said:Sorry, but it is clear that you have no idea what goes on in a dial-up
connection. Forget the start and stop bits. Unless you have disabled
error correction, these bits are stripped out and discarded by the
modem.
To see that this is indeed the case, just look at my test
results. These test data were obtained by using HyperTerminal to
transmit a plain text file directly between two modems connected via a
short piece of phone cable (a perfect noiseless line).
In a non-EC connection, the two modems transfer raw data including
start and stop bits, as follows.
10 bits 10 bits 10 bits
PC #1 <------> Modem #1 <<-------->> Modem #2 <------> PC #2
DTE DCE DTE
An EC connection (without data compression) works as follows.
10 bits 8 bits 10 bits
PC #1 <------> Modem #1 <<-------->> Modem #2 <------> PC #2
DTE DCE DTE
In the latter case there are no start and stop bits. Instead the data
are grouped into packets (or blocks) of 64, 128, or 256 bytes, plus a
small number (6 or 7?) of additional header bytes. AFAIK, the header
includes sync bytes for setting up the clocking of subsequent data
bytes, and a checksum (CRC) for verifying the integrity of the packet.
When you connect to the Internet, you are using PPP. This protocol
adds quite a lot of overhead, so that the download rate reported by
your browser is a lot less than the actual transfer rate of your
modem. That is, your browser counts only the file data, not file data
*plus* PPP overhead. To see the actual download rate you need an app
such as System Monitor (sysmon.exe) which ships with Win9x. For a
46667bps connection, Sysmon reports a download rate of 5.3KB/s for
compressed files, while my browser may show only 4.8KB/s, say.
half_pint said:Actually I now accept that the modems (to modem) may not be using
stop and start bits.
I am sorry I am really not to sure what you are on about,
what have start and stop bits got to do with error correction?
Nothing whatsoever as far as I am concerned.
You will have to explain yourself better as I am not sure
what error correction you are on about.
As far as I can see there is no option to 'disable error
connection' in internet options,
you can forget Hyperterminal
because I don't use it (as far as I am aware) and neither does the OP.
So i am not really to sure what you are getting at because when I
look at my modem properties is says 8 data, 1 stop bit and no
parity bit (I presume the start bit is taken for granted).
Because it says this I have to believe what I see and take it
as fact as that is what the modem is doing.
If it was configured otherwise I am sure it would say so.
If you can show me a link which clearly shows that these value
are not used I would like to see it.
As I said your Hyper Terminal experiments are not really relevant
as far as I can see.
The utility I mentioned in this thread shows me receiving data at 6.2KBs
Incidently the utility I am using shows me receiveing data at 7.3 KBs
which is faster than a 56k modem is alledgedly capable!!
Actually I now accept that the modems (to modem) may not be using
stop and start bits. However having said that I doubt his problem is
down to the telephone line which is probably capable of transmitting
8 mega bits per second, I mean it would be a bit odd if it was only
managing 0.7% of its capacity, and even odder is several people
miraculously also received a similar degradation?
Franc Zabkar said:Exactly. Yet it is you who has introduced irrelevant start and stop
bits to "explain" modem-to-modem transfer rates. What I am saying is
that the only time modems transfer start and stop bits between each
other is on those very rare occasions they have been unable to
negotiate an error corrected link. At all other times these bits are
discarded.
Modems communicate between each other using an error correction
protocol, either V42 LAPM or MNP 4. They assemble data in blocks or
packets, usually 64, 128, or 256 bytes in size. The receiving modem
checks the integrity of the block by validating a checksum (CRC). If
the CRC does not match, then the receiving modem requests that the
faulty block be retransmitted by the sending modem.
In Win9x, go to My Computer -> Dial-Up Networking -> right click your
ISP -> Properties -> General -> Configure -> Connection -> Advanced.
There is a check box called "Use error control". You also have the
option to "compress data".
IMHO, the only way to get meaningful, reproducible results is to use a
comms app such as HyperTerminal, in the way that I have outlined
above.
What you are seeing is the configuration of your COM port, ie the
*DTE* interface, not the *DCE*. The PC sends data bytes out the COM
port, framed with start and stop bits and optional parity, at a DTE
speed of 115200bps, say. However, the two modems discard these extra
bits when communicating between themselves. You need to learn the
difference between the terms DTE and DCE.
DUN has no idea what transpires between the two modems. Such things
are mostly beyond its control.
They *are* used, but only by the DTE, not the DCE. If you compare the
first two lines of my test results, all should be clear (see the
diagrams below). I suggest you also lurk at comp.dcom.modems for a
while.
On the contrary, it is your arbitrary, uncontrolled Internet based
"testing" which is irrelevant. In my test setup I have control over
*all* variables, namely the connection speed (constant 33600bps), line
quality (perfect), and data type (raw text with no PPP overhead).
There are also no potential issues with network congestion.
That test is pointless from the point of view of this discussion. The
figure of 6.2KBs does not reflect the speed of your connection,
instead it shows how well your modem is able to compress the test
data. You need to monitor the download rates for *incompressible*
data, eg ZIPs.
FWIW, my internal modem can achieve download rates as high as 21KB/s
when tranferring highly compressible data such as some newsgroup
headers.
Franc Zabkar said:Think "data compression".
AFAIK, the bandwidth of a voice line (~4KHz) is restricted by the line
card in the telephone exchange, and by regulations. ADSL is another
matter. I don't know much about it as yet, but hopefully that will
change when I switch over during the coming week.
On our standard 56k dialup modem (in the UK, this is), we only get ~38k when
we connect. Are there any ways to correct this? I've considered calling
the phone company and asking them to turn up the gain, but I've also been
told that this wouldn't make any difference.
Drivers etc are all up to date and this is happening on 2 computers in our
house, so does anyone have any ideas?
Cheers for your help.
<Barney />
Gary D. said:I previously suffered with an almost identical problem, although apart
from slow connection speeds my main issue was random disconnections.
My ISP suggested a few things, including contacting my phone company
and asking them to increase the line gain from the standard 1 to the
maximum of 3.
However, this is fine with a clean line but with a noisy line will
only increase the noise and also increase the likelihood of random
disconnections, so first get the noise problem sorted out.
When my phone company examined my nearest public routing/junction
cabinet the engineer discovered that during heavy rainfall the water
had been seeping into the cabinet and had gradually eroded some of the
contacts; these were replaced and the line gain was increased to its
maximum and now I connect at 50667bps for 90% of the time (49333bps
the rest of the time). The connection rate is also now much more
stable with only an occasional random disconnection.
I would imagine any restrictions were simply filters are the exchange.
You appear to have snipped out the bit which proves I was right all along?
I introduded the stop and start bit because they are relevant
to the comms.
Obviously even if these bits are stripped out by the modem it
has no effect on the overall speed because the modem still
has to wait to receive these bits so it might as well transmit them
because it will have bugger all else to do whilst it is waiting.
You sure about that?
what about a 1 byte keystroke?
Those obtions are not available on my PC, the boxes are there but
they are inaccessible.
I know the differemce and it says modem properties not DTE so you need
to learn to read what is written.
It don't think it really matters as the comms speed it determinded by the
DTE
speed not the modem.
However you are not testing the right patient so to speak,
you are testing a different patient and assuming the real patient
has the same symptoms you found on the proxy patient.
A rather foolish way to diagnose a problem.
The data I am using is compressed data eg jpegs.
well mine is showing 36KBs ...
... but I would rather see what
rate bits are being transmitted at as anything else is pretty meaning less.
Gary D. said:I previously suffered with an almost identical problem, although apart
from slow connection speeds my main issue was random disconnections.
My ISP suggested a few things, including contacting my phone company
and asking them to increase the line gain from the standard 1 to the
maximum of 3.
However, this is fine with a clean line but with a noisy line will
only increase the noise and also increase the likelihood of random
disconnections, so first get the noise problem sorted out.
When my phone company examined my nearest public routing/junction
cabinet the engineer discovered that during heavy rainfall the water
had been seeping into the cabinet and had gradually eroded some of the
contacts; these were replaced and the line gain was increased to its
maximum and now I connect at 50667bps for 90% of the time (49333bps
the rest of the time). The connection rate is also now much more
stable with only an occasional random disconnection.
It might also be worth him checking inside the telephone sockets inside
his home as water and condensation can get inside and corode the contacts,
or make false connections across them.
I had this happen to me, the engineer chased a noisy line down to corosion
inside one of the sockets.
I live in the UK by the way.
Don't know if any of this info helps?
PS. I'm using Windows98 with an internal "Conexant HCF V90 56K Data
Fax RTAD PCI Modem". I've also used a "Rockwell 56K ACF II
Fax+Data+Voice Modem" (external) with so-so results.
Just remembered something else!
I have been informed that if there are several communications devices
connected sharing the same line (eg. PC, phone, fax) then by
disconnecting all others when your PC's modem is being used, this can
apparently improve connection speeds.
Try it!
I don't know the exact reason why this should work, but I'm guessing
it may be something to do with the other devices having to process
(and ignore) the noise generated by the modem signal.
Gary D. said:~38k
when
It might also be worth him checking inside the telephone sockets inside
his home as water and condensation can get inside and corode the contacts,
or make false connections across them.
I had this happen to me, the engineer chased a noisy line down to corosion
inside one of the sockets.
Just remembered something else!
I have been informed that if there are several communications devices
connected sharing the same line (eg. PC, phone, fax) then by
disconnecting all others when your PC's modem is being used, this can
apparently improve connection speeds.
Try it!
I don't know the exact reason why this should work, but I'm guessing
it may be something to do with the other devices having to process
(and ignore) the noise generated by the modem signal.
When I asked BT about my bad line they always asked how many
devices were connected to it, I suppose the less things you have connected
the less chance you have of one of them being faulty? Also I think a battery
at the exchange provides power for the phone(s)? maybe there is a limit as
to
how many devices you can have connected?
Gary D. said:It might also be worth him checking inside the telephone sockets inside
his home as water and condensation can get inside and corode the contacts,
or make false connections across them.
I had this happen to me, the engineer chased a noisy line down to corosion
inside one of the sockets.
Just remembered something else!
I have been informed that if there are several communications devices
connected sharing the same line (eg. PC, phone, fax) then by
disconnecting all others when your PC's modem is being used, this can
apparently improve connection speeds.
Try it!
I don't know the exact reason why this should work, but I'm guessing
it may be something to do with the other devices having to process
(and ignore) the noise generated by the modem signal.
Living in a remote corner of India, I thought almost everyone in the
more advanced countries would be using at least >128Kb/s cable/DSL.
Here's my 2 cents worth from an end-user's POV.
My phone company is also my ISP and I happen to live less than 200
metres from their building, though the actual line length is probably
about twice that. I've used several different modems and never
bothered to use anything other than the generic Windows drivers. Max
connection speed setting is limited to 115K, and I often manage to
connect at that speed, though it's sometimes at 33k or even lower.
The connecting speed doesn't seem to greatly determine the actual
download speed. I may connect at 115k and still have to wait for ages
for web pages to open, or connect at 33k another time and have the
same website move at relatively blazing speeds.
File download speeds as shown by IE or Download Accelerator vary from
less than 0.1 kiloBYTES/s to about 4.5kB/s on a very good day. These
are sustained speeds, not burst speeds which sometimes shoot up to
20kB/s at the start of a file download. The primarily cause of the
large variation in speed does not seem to be the state of my phone
line (though this must certainly be a factor), but rather from the
server since friends in other parts of the town usually experience
corresponding variations at the same time.
Another factor that took some time to track down was one particular
computer's PSU. It caused connection problems, random disconnections,
etc., but ONLY when my phone line was unusually noisy. Replacing the
PSU cured it completely. It never posed any problem in off-line
computing.
BTW, one of the "advantages" of living in a less developed region is
that the phone company doesn't seem to mind if I repair or tweak my
own line. Twice in the past 10 years or so, I've bought phone cable
from a store and completely replaced the overhead lines from their
junction box. Fibre-optic has not completely replaced metal here.