To: rsgb_lf_group@blacksheep.org, SpectrumLabUsers@yahoogroups.com
References: <56BDED97.5060308@freenet.de> <56BE0364.6070503@gmx.net>
From: =?UTF-8?Q?Wolfgang_B=c3=bcscher?= <dl4yhf@freenet.de>
Message-ID: <56BE1134.7060205@freenet.de>
Date: Fri, 12 Feb 2016 18:07:00 +0100
User-Agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:38.0) Gecko/20100101
 Thunderbird/38.5.1
MIME-Version: 1.0
In-Reply-To: <56BE0364.6070503@gmx.net>
Subject: Re: LF: PIC-based GPSDO with A/D converter / supported serial
 baudrates ?
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Status: RO

Hi Tobias,

Thanks for the information. 12 Mbaud would be fine for a 'real' SDR 
already, and far more than a PIC with 40 MHz oscillator frequency ( -> 
10 MHz 'cycle' clock) needs. To keep the project homebrew-friendly the 
plan is to use an off-the-shelf adapter (like the ones FTDI call 'USB 
RS232 Cables' even though they are much more than just a cable..).

The 'Prolific' seems unable to read his own transmitted signal at the 
higher bitrates. I tried it with my own software (WinPic, which contains 
a terminal emulator for debugging), where I can enter *any* numeric 
baudrate. But windows (or the driver) never complains, so there is no 
easy way to check it.

In the meantime, 460800 bits/second seem to work, which should be 
sufficient for approx. 22 kSamples with one channel, and 16 (averaged) 
bits per sample. Not a replacement for a HiFi soundcard (and not "I/Q" 
at all), but with a similar principle as recently suggested by G4JNT 
(using only one channel but a special sampling sequenze) may work just 
as well.
I will find out soon what the PIC16F's max ADC sampling rate of 75 
kS/sec can do.. and if the XC8 compiler permits a bit of signal 
pre-processing (or if I really need to switch to a dsPIC. I hope not..)

A block diagram of the GPSDO (as it looks like at this very moment)
is at http://www.qsl.net/dl4yhf/gpsdo/gpsdo_pic_main.c  .

Should anyone here consider building the GPSDO: please be patient - the 
circuit (contained in the sourcecode as 'ascii art') may be subject to 
change. I may even decide to replace the PIC16F1783 by a dsPIC30F3013, 
which unfortunately has an incompatible pinout (despite the same SDIP-28 
footprint).


73,
   Wolf DL4YHF .


Am 12.02.2016 um 17:08 schrieb Tobias DG3LV:
> Hi Wolf !
>
> Please have a look at :
> http://www.ftdichip.com/Products/ICs/FT4232H.htm
>
> it says : "RS232/RS422/RS485 UART Transfer Data Rate up to 12Mbaud. 
> (RS232 Data Rate limited by external level shifter)."
>
> The datasheet
> http://www.ftdichip.com/Support/Documents/DataSheets/ICs/DS_FT4232H.pdf claims 
> :
> "Data Transfer rate: The FT4232H supports a data transfer rate up to 
> 12 Mbit/s when configured as an RS232/RS422/RS485 UART interface. 
> Please note the FT4232H does not support the baud rates of 7 Mbaud 9 
> Mbaud, 10 Mbaud and 11 Mbaud."
>
> and :
>
> "The Baud Rate Generators provides a x16 or a x10 clock input
> to the UART’s from a 120MHz reference clock and consists of a 14 bit 
> pre-scaler and 4 register bits which provide fine tuning of the baud 
> rate (used to divide by a number plus a fraction). This determines the
> Baud Rate of the UART which is programmable from 183 baud to 12 
> million baud."
>
> If you can omit the (RS232-) level shifter and just use the 5 Volt 
> TTL-Level in/output directly, then higher baudrates than 115200 should 
> be possible at not too long cable-distances. (if possible, use 
> RS422/RS485 differential signalling) At the QRL we did tests with this 
> chip at up to 4 MBit some time ago with some success using the VCP 
> driver. (There may be other chips from FTDI having a similar 
> performance, we just selected this one.)
>
> The FT4232H chip is, too, available as a (not-too-expensive, €28.71) 
> mini-module at their webshop. (The more prominent chips from FTDI may 
> be a target to fakes at the "usual market platforms")
>
> The Windows(/Linux)-program on the PC has to be able to configure 
> these baudrates. Standard Terminal-programs will support up to 115200 
> Baud. But self-created programs could go well beyond this (using the 
> WIN32 -API).
>
> 73 de dg3lv Tobias
>
> Am 12.02.2016 um 15:35 schrieb Wolfgang Büscher:
>> Greetings all,
>>
>> The PIC-based GPSDO I am currently building (using a PIC16F1783, but
>> maybe in the end it will turn out to be a dsPIC30F3013) is happily
>> synchronizing the surplus 10 MHz VCOCXO for over a week now, without a
>> single hiccup.
>> So far, so good... I'm quite happy with the very low phase noise from
>> the control loop (which uses two integrator to bring back the phase
>> error to zero, should there ever be a 'hiccup' which has not happened 
>> yet).
>> The next plan is to use the PIC's internal 12-bit ADC as a replacement
>> for the soundcard, and possibly the PIC's internal PSMC as a simple
>> downconverter ( programmable switch mode controller, can supply 40 MHz /
>> N or 20 MHz / (N + M/16) with the fractional frequency adjustment ).
>> To keep it simple, without an extra DDS, frequencies which the GPSDO can
>> produce 'directly' (using the PSMC) are for example:
>>
>> 476190.47619048 Hz
>> 475498.33887043 Hz
>> 474806.20155039 Hz
>> 474114.06423034 Hz
>> 473421.92691030 Hz
>> 472729.78959025 Hz
>> 472037.65227021 Hz
>>
>>   (for the curious reader, the above frequenies are taken from the 10
>> MHz OCXO, multiplied by the PIC's internal PLL to 40 MHz, then divided
>> down by the PSMC as described in the PIC16F1782/3 datasheet,
>> DS40001579E, on page 220. One could even use the PSMC as a crude BPSK
>> modulator, but that's off topic for the moment).
>>
>> Since I don't want to develop an own USB driver for the digitized input,
>> the interface shall use the UART (RS232 or RS422, this is up to the
>> re-builder). The question is which serial baudrates (bits per second)
>> are supported by the common "RS232" / USB adapters. They all support
>> 115200 bit/sec but that's a bit low (if most of the digital
>> downconversion shall be performed by the PC, due to the stepwidth of the
>> PSMC output), so higher serial baudrates for the transmission of analog
>> samples (from PIC to PC) are desirable. The question is WHICH higher
>> baudrates are available, and is there a simple way on a PC (Linux or
>> Windows, shouldn't matter).
>> Must these all be multiples of 115200 ? So far I could only test it with
>> the already mentioned 'Prolific' adapter, but I can see no way how to
>> enumerate the baudrates which the adapter *really* supports.
>> Trial-and-error experiments with various bitrates only produced garbage
>> data. The rise and fall times at the RS232-TXD output are about 800 ns
>> which is surprisingly steep so I guess the level converter itself cannot
>> be the problem. Any ideas / recommendations ?
>> Buy a 'good' FTDI adapter (instead of the Prolific), and if so, which
>> one exactly ?
>> If you have one of those, can you switch it to 1 MBit/second, hook it up
>> to a scope, send a few characters from a terminal program and check the
>> bit timing + rise & fall times ?
>>
>> 73,  Wolf DL4YHF