[pskmail] Re: 300 Baud Rate in the USA
- From: John Douyere <vk2eta@xxxxxxxxx>
- To: pskmail@xxxxxxxxxxxxx
- Date: Wed, 25 Jan 2012 10:17:53 +1100
Hello Bernard,
Thank you for your reply.
I agree with all you said in your email.
My question was specifically directed (and obviously I must have
introduced some confusion there) at the speed of the exchange.
I can spend time making Pskmail faster with new modes or add new
features or new platforms.
But it is a trade-off, hence my question: what is the practical value
of let's say doubling or quadrupling the throughput speed?
These modes would be wider and less sensitive as these go hand in hand
with higher data speed for a given HF link.
73, John
On Wed, Jan 25, 2012 at 7:20 AM, Bernard Dekok <kc9sgv@xxxxxxxxx> wrote:
> Hi John,
>
> Thanks for the excellent explanation about baud rate in the modern
> environment.
>
> "What for" should this be done ? Hmm......
>
> It is Ham radio. Maybe the envelope is not exactly pushed here, as I am sure
> there are many military standard and commercial standard data protocols that
> may do this already.
> For a price.
>
> The cash-strapped camper, hiker, hunter, fisherman, boater (and their
> families), might applaud the idea that Ham radio can now send emails via HF
> for FREE.
> I am sure all of us are having fun with it.
>
> Now you can buy a Spot Tracker for US $200 and a $150/year service contract
> to track your whereabouts on some parts of Earth and to send APRS-style 160
> character text messages.....These are great safety devices if you are in the
> satellite footprints. They save hundreds of lives per year already.
>
> For audio and high data transfer satellite phone connections, there are
> numerous services like Iridium etc.
> Last time I checked, the least expensive plans were about $1500 for a sat
> phone and about $1.50/minute for audio and about $15/minute for slow data.
>
> See
> here: http://www.satphonestore.com/satellite-phones/satellite-phone-hh.html?manufacturer=9
>
> More research on this topic is required. Bottom line: Satellite
> communications is still pricey.
>
> KC9SGV
>
>
>
>
> On Mon, Jan 23, 2012 at 8:26 PM, John Douyere <vk2eta@xxxxxxxxx> wrote:
>>
>> Maybe a little explanation about the baud rate limits.
>>
>> Going back some time (most likely when RTTY was THE mode), bits per
>> second and baud rate were the same. Let me explain:
>>
>> The baud rate measures how many times the signal changes
>> characteristic (frequency, phase, amplitude or any combination of the
>> above) in a second.
>>
>> This limit in the USA is 300 changes per second. For RTTY, where there
>> is a mark and a space used in alternance, one change (one baud)
>> equates to one bit of information (a 0 or a 1) so in that case baud
>> rate = bit rate per second.
>>
>> What has happened in time of course if that some clever person asked,
>> why can't I send more information through the same pipe?
>>
>> So take for example the MFSK16 mode: in that case we change the
>> frequency of the signal every 64 milliseconds giving a baud rate of
>> 15.625 baud (1 / 0.064 = 15.625).
>>
>> Since the signal can change to any of 16 different defined
>> frequencies, it is interpreted as a number between 0 and 15,
>> equivalent to 4 bits of data (2 power 4 = 16).
>>
>> Therefore the bits per second are 15.625 * 4 = 62.5 bits per second,
>> much higher than the baud rate.
>>
>> For a PSK type of signal where the changes between symbols are defined
>> by phase, if we want to push more bits per second than baud we have to
>> go for less than 180 phase changes to say 90 degree changes (or
>> multiples of, as in 0 or 90 or 180 or 270 = 4 combinations = 2 bits).
>>
>> That way we would have 2 bits per baud, double the data rate at the
>> same rate of signal change or baud rate.
>>
>> This is what QPSK does, and its raw data speed is double the rate of
>> standard PSK.
>>
>> So why is it given as the same data rate as PSK? It is slowed back
>> down by half by the Forward Error Correction that is added on top,
>> where each data bit is sent twice, resulting in the same net rate as
>> standard PSK.
>>
>> So we could have 8PSK where the changes are multiples of 45 degrees,
>> carrying 3 bits per symbol etc...
>>
>> The only issue becomes the bandwidth, the sensitivity and the
>> resilience to interferences. A swing of 180 degrees in PSK is more
>> likely to be correctly decoded than a swing of 45 degrees when you add
>> the effect of the propagation and QRN/QRM. Also as the signal
>> bandwidth increases the sensitivity reduces, or alternatively, the
>> power required for the same reliability is increased.
>>
>> So there is a practical limit from that aspect.
>>
>> What could we do in practice with what we have (sound card based
>> systems are not very precise devices in terms of timing)?
>>
>> In my opinion we could either:
>>
>>
>> A. Go wider and have multiple PSK250/PSK250R modulations (carriers) in
>> parallel.
>>
>> B. Develop 8PSK, 16PSK of modes and improve the FEC scheme.
>>
>> C. Combine the two above (in principle what WINMOR does).
>>
>> D. Get the 300 baud limit of the FCC removed and use the PSK500 mode.
>>
>> E. Combine all of the above :)
>>
>> As exiting as these developments may be, I always come back to the
>> "what for" question.
>>
>> I haven't yet seen a case, backed by data, justifying the effort, but
>> I am open to suggestions.
>>
>> Until now I have seen more value in the addition of features, and
>> Pskmail has quite a few compared to the alternatives, and the
>> expansion across platforms (Windows, Linux, Android) than in breaking
>> speed records, but as I said, I am open to revisit that view.
>>
>> Just my opinion, not representative of the development team.
>>
>> Hope this helps,
>>
>> 73, John
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> On Tue, Jan 24, 2012 at 10:50 AM, David Kleber <kb3fxi@xxxxxxxxx> wrote:
>> > On a recent peer to peer test with winmor 1600 on a good channel, it
>> > took
>> > about 10 minutes to send around 30k of data with few if any retries. It
>> > takes about 10 minutes to send a 6k file with MT63 1k long. So, from my
>> > very
>> > unofficial observations, Winmor is 5 times faster than MT63 1k long and
>> > should be about 500 wpm, or about the same speed as BPSK500, but with
>> > much
>> > better accuracy.
>> >
>> > -Dave, KB3FXI
>> >
>> >
>> > ________________________________
>> > From: Bernard Dekok <kc9sgv@xxxxxxxxx>
>> > To: pskmail@xxxxxxxxxxxxx
>> > Sent: Monday, January 23, 2012 5:51 PM
>> > Subject: [pskmail] Re: 300 Baud Rate in the USA
>> >
>> > Dave,
>> >
>> > Baud rate can't be WPM / 1.2
>> > Winmor 8 Car 16PSK is 94 baud, at 3285 WPM on 1600 HZ bandwidth.
>> > See the Excel spreadsheet on this thread.
>> > Yep, 3285 words per minute.
>> > Try doing THAT on CW.....=)
>> >
>> > I vote for the Winmor virtual TNC as another mode option for PSKMail....
>> >
>> > Bernie,
>> > KC9SGV
>> >
>> >
>> > On Mon, Jan 23, 2012 at 3:06 PM, David Kleber <kb3fxi@xxxxxxxxx> wrote:
>> >
>> > I think I recall seeing someone post somewhere that BPSK500 goes over
>> > the US
>> > throughput limit, but I've never seen what appears to be a well detailed
>> > explanation. Also, there seems to be varying opinions on how to
>> > calculate
>> > baud rates from wpm (wpm divided by 1.2?). It's all quite confusing.
>> >
>> > I wish we could just go to regulation by bandwidth with a separate
>> > playground for unattended ops. The robots in any bandwidth need to have
>> > their own allocations which makes sense for anyone on either side of the
>> > fence (whether you hate the robots or love them).
>> >
>> > We really need to get rid of the throughput limitation, which
>> > contradicts
>> > one of the key reasons for our amateur radio frequency allocation which
>> > is
>> > the advancement of the art (and technology).
>> >
>> > BTW, BPSK500 leaves a bit to be desired unless you're on a near full
>> > quieting path.
>> >
>> > -Dave, KB3FXI
>> >
>> >
>> > ________________________________
>> > From: Bernard Dekok <kc9sgv@xxxxxxxxx>
>> > To: pskmail@xxxxxxxxxxxxx
>> > Sent: Monday, January 23, 2012 3:49 PM
>> > Subject: [pskmail] 300 Baud Rate in the USA
>> >
>> > Ok,
>> > So I read up a bit on this here in the PSKMail archives.
>> > 300 Baud max on the HF bands in the USA.
>> >
>> > So what is the PSKMail baud rate at 500 HZ ?
>> > Why is PSK500 not allowed in the USA, when Winmor can use the 500 HZ
>> > mode
>> > unattended ?
>> > Winmor clients can indeed use the 1600 HZ mode if the station is
>> > attended.
>> > Some USA Winmor servers (RMS's), even run at 1600 HZ, in "attended"
>> > mode.
>> >
>> > What is interesting is that the Winmor baud rate is only 93 baud on 1600
>> > HZ.
>> > See here and download "Winmor Rate":
>> > http://www.winlink.org/WINMOR
>> >
>> > Could we run a PSKMail PSK500 server in "attended" mode in the USA ?
>> > How about "unattended", but under control operator/sysop control via
>> > remote
>> > ops (say by VHF or Iphone/Echolink) from the lawnmower in the yard while
>> > indeed mowing the lawn ?
>> >
>> > We need more and faster PSKMail servers in the USA.
>> >
>> > Bernie,
>> > KC9SGV
>> > Chicago.
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>>
>
Other related posts:
