The Radio Broadcaster's Dream Mini Rack
8 comments
·April 3, 2025Nextgrid
geerlingguy
Indeed, even many of the devices here could be running on a Pi internally... or are running Linux on some other Arm SoC or a suite of microcontrollers.
One of the new audio processors (full size rack) actually runs a small cluster of CM4s for local audio processing, a remote control system, and something else (can't remember), but the convergence from all these discrete specialized boxes to one or a few general PCs is slow to happen.
Some of that is down to proprietary stuff, but a lot is inertia, and being allergic to less-tested solutions.
The Inovonics box could be replaced by Stereo Tool [1] on a Pi—and indeed it is in some lower budget installs!
A lot of radio engineers are also hams and so are willing to experiment sometimes, especially in lower power stations where the stakes aren't as high. But if you're dealing with one of the "big" stations or ownership that has the IBM mentality (nobody got fired for choosing IBM), they tend to stick to the tried and tested audio chain, with more robust equipment with a long history of support.
cf100clunk
The authorities are very picky about the bench-tested qualities of radio and television gear in professional use, and specialization of capabilities across different brands leads to best of breed gear options. It is a very serious business and I have the highest regard for the radio and television engineers who oversee such gear.
For outsiders it is a bit like thinking about how an RPI5 could theoretically run medical gear in an operating theatre or control the just-in-time processes of a major factory, but would regulators and insurance providers go for it? For amateur broadcasting, sure, an RPI can have fun uses like this:
0xbadc0de5
Aside: As someone who also has a great relationship with his father, I always notice this sort of thing in others.
Great work, team.
JKCalhoun
While PoE is nice, I suspect most of those wall-warts output something like 9V or perhaps even 5V. If there are matching wall-warts, a small switching power supply of the same voltage (and combined current) would be a clean replacement.
The point about a UPS is a good one though. There is certainly an efficiency loss to step up to 110V, step back down to 5V.
Having fitted out an RV though, it was nice to find lots of 12V appliances — they could be run into a small 12V fuse block and the fuse block wired directly to a 12V "house" battery.
cf100clunk
I like the compactness, and since it seems to be North American gear I'm wondering if additional gear like an HD Radio (IBOC) modulator-injector such as a Nautel HD MultiCast+ could be part of the chain? I wish there was progress on an open source IBOC exporter but it doesn't seem to be a thing and its waaay outside of my capabilities.
geerlingguy
HD Radio is still (annoyingly) covered by patents in a way that open source efforts like nrsc5 still can't be integrated into any OSS radio tooling for fear of litigation.
And from my own IANAL research, it seems iffy whether the initial set of patents which is expiring soon are the ones that will open it up... or if any of the many follow up set of patents can keep it out of easy reverse-engineered access :(
adamgordonbell
I think but am not sure that the goal here is to deliver reliable audio to a transmitter at a remote site from a radio station.
I find it curious that so many separate components are used (that presumably aren't cheap, as they target a low-volume, pretty niche industry).
It seems to me that most of these can be replaced by any computer with an analog audio output (and since this is Jeff Geerling we're talking about, it can be a Raspberry PI)?
Maybe put two and use the "silence detector" box to automatically switch between them in case one goes down?
In fact, I wonder if a custom board can be designed with 2 separate stages (each powered by a compute module and DAC) and some circuitry to do the silence detection and switching, with some GPIO lines going back to each to inform them of which one is being the chosen one for the final audio output. This would effectively condense this whole thing into a single box. But not being a broadcast engineer, I'm just talking out of my ass here, maybe there's a good reason they're doing things they way they do.