We appreciate your thoughts on making more USB sound fobs usable with ASL. And I don’t want to dampen your enthusiasm about it.
In a perfect scenario that would be the desired choice and outcome for ASL.
With Allstarlink there is a lot more than just getting sound to play out of USB sound fob.
Two items that come to mind are the stability issues and timing issues not heard by the human ear that are needed by app_rpt software to do certain functions.
So, So, So much more going on with the app_rpt software than one would think.
I just want to provide you an example of the difficulty in even getting two so called compatible sound chips the CM119A verses the CM119B working correctly with app_rpt.
These two chips are made by the same manufacture! Even the manufacture stated that these two chips were compatible with each other! And we quickly discovered this was not the case!
Please see this previous post explain some of the issues posted a while ago regarding the work and man hours spent just to figure out these differences.
Now times these differences by 100 of the new problems created when dealing with another manufacture of a USB sound fob that has different specifications for their sound chip?
Do you get how this could be so overwhelming that it would practically be impossible to deal with a bunch of different sound fobs?
From: Derek Chauran via AllStarLink Discussion Groups [mailto:firstname.lastname@example.org]
Sent: Thursday, July 02, 2020 9:26 PM
Subject: [AllStarLink Discussion Groups] [App_rpt-users] RPi Gpio
I’m going to have to disagree that USB and serial fobs are a moving target. Neither require direct support of a specific device, as both have robust and stable abstraction layers. That’s precisely why they are so popular for makers/hams/etc… tty serial hasn’t changed in any major way in linux since… the early 1990s? And Alsa has been around since 1998. ASL doesn’t even have the burden of supporting multiple operating systems that complicates these hardware abstractions.
As to the cost - If I could use any old USB fob ($3 for $10) and the built in pi serial interface I could build an entire duplex node for less than the cost of a URI, and still have overhead on that pi to add a second node (about $25), an RTL-TCP server ($20-30) and probably more.
This is also a bit of a privileged point of view. A lot of our community where I live struggle to afford the cost of a decent radio, or a QTH that allows them a usable antenna in an RF friendly environment. For them that $50 or $70 could be the deciding factor for a friend building them a node with the pi 3 they just replaced with a 4b, or the difference between saving for a few months to build a node and deciding it’s too expensive. Fortunately, we have echolink to offer those people, but echolink has numerous disadvantages, which I am sure are well known here.
Visit Topic or reply to this email to respond.
Having already been down this path, there are more challenges than you realize. At some point you have to standardize on supported solutions----or go crazy trying to support a gazillion custom one-off systems that end up with odd problems.
I guess the question comes down to are you an end-user who wants to build a few nodes and get on AllStar, or are you wanting to manufacture a new product. If an end-user, once you start adding up all the time and costs, $50 isn’t expensive for an adapter. If you’re building a new product, just realize that very low-end USB sound FOBs and serial adapters are a rapidly moving target…This is why all the US hams building adapters ultimately do their own thing—they had to have consistency and control quality. The same goes for cheap/clone USB serial adapters. While these adapters may work fine for PTT (using DTR or RTS), once you try to obtain COS signalling, you find that reading CTS or DSR doesn’t have the timing repeatability that is required. So, you end up with chopped off words, etc.
73, David KB4FXC
All of these are fine devices for someone who just wants to throw money at a node. Personally, I like to build stuff and do it on the cheap.
The RA adapters are $25 minimum, with no case, for something that has the same functionality as the USB sound card. The one person working on the Dinah/Brian is busy with his Shari backorders right now (really nice guy, quite open to feedback!), not to mention that is an expensive option, TechByGeorge - really nuce stuff - $300 for a node, or $50 for essentially a USB sound card with a parallel port that I don’t need - no thanks, and the URI is $70 for the same thing, but in a case.
Meanwhile, I can interface with any of my radios from FLDIGI, HRD, WSJT-X, DXLabs Commander, etc… using a $3 USB audio controller (if using a pi), a cable or 2 (most of these can be scavenged), and a free GPIO serial port or a $5 USB serial adapter (if using a PC). Technically, I can do this all with no soldering BTW.
It just really doesn’t strike me as a good idea to have precisely 1 supported hardware configuration, which is third party, and which requires either very good soldering skills, or require people purchase expensive premade adapters, when there are literally free options available on the device that most people are using to deploy new nodes, are much easier to setup, and are supported by just about every other ham software out there. I also am not aware of any radio manufacturer who publishes an interface to their radios that involves soldering pins on a tiny SMD component, but plenty publish schematics for using a serial port and and audio connection.
While CM1xyz boards from China are slower to obtain RIGHT NOW due to Covid-19, there are a multitude of adapters available from many sources in the USA. Here are links to several sources. The CM1xyz is not in going away (or getting more expensive) anytime soon:
73, David KB4FXC
Thanks Marshall - I get that the devs have a limited amount of time to spend writing features like this, and I very much understand that they’re never as simple as one would expect (I’m a software engineer in a different ecosystem). I guess the main issue that I have is that I’ve dug up threads with this request going back years, and the answer always seems to be either indifference or “deal with it”. The hardware cost and/or complexity is a major gating factor to people using ASL. I am definitely not alone in this thinking.
At any rate, rather than GPIO, I will look into how to make this work with the linux serial interface. That seems like the best approach, as the rpi gpio serial interface can be setup to use GPIO pins for RTS (ptt) and CTS (cos) in addition to the communication interface. I’ve read the old LPT port info, but it doesn’t seem to apply. Someone else (who shares in my frustration) also found a very old patch where someone implemented alsa+serial, so that might have something I can build from.
You do some great ideas and I agree with you on the COS and PTT.
It would certainly make things easier for hams to interface radios to soundcards.
I was kind of trying to be funny when I made mention of the adding these features to the open source project.
What I was really getting at is, there is a lot of work involved with making a system like Allstarlink work.
Sometimes it is not very simple to add features.
There are only a few folks with the programming skills necessary to write or modify the open source to make an alternate mechanism for COS and PTT.
(Until now, no one has offered to step forward and take on the project of an alternate mechanism for COS and PTT.)
I could be wrong, but at some point in time there was some provision for COS and PTT using the parallel port of a PC.
We all enjoy what is offered for Allstarlink for FREE, there is no charge for the software or service.
Sometimes we just have to use what is available, even if it is not the easiest to deal with.
Marshall - ke6pcv
··· (click for more details)
Visit Topic or reply to this email to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, click here.