Apache Labs Community Forum

Hi All,
Does anyone know if the new G1 will run P2 with Thetis?
Frank WA1GFZ

'support' being mentioned, who will support the software side of G1 things? Was it a group of hams that designed it and made the FPGA code open source and is that accessible somewhere?
Or is the G1 a proprietary thing, to be fully supported by A-L?
Reading the announcement it states that the G1 is compatible with a control program like Pihpsdr. Yet, the latest source code of Pihpsdr doesn't 'know' a G1.
Wouldn't the G1 deserve its own section on this forum, it having little commonality other than its name and brand?
The G2 section is already quite complex with G2's many configuration options of stand alone vs external control, with/without control front, old and new control front, with and without CM5 upgrade.
And soon a v3 Saturn PCB upgrade being introduced. The G1 may be a source of confusion in a G2 dedicated section.

ct1iqi wrote: Mon Oct 20, 2025 1:11 pm 'support' being mentioned, who will support the software side of G1 things? Was it a group of hams that designed it and made the FPGA code open source and is that accessible somewhere?
Or is the G1 a proprietary thing, to be fully supported by A-L?

Upon further reflection, this is an excellent question. And I'm going to retract (delete) my previous comment about it being delivered with P2.

It is conceivable that Apache themselves might create their own P1-variant firmware for it, using the existing 8000 P1 firmware as a basis. This would not be terribly difficult. Achieving timing closure on the P1 firmware is straightforward and the architecture and interfaces of the G1 are essentially identical to the 8000 (and all Apache SDR boards are not so very different from the original Hermes design). Such firmware would be a derivative work, and should therefore be published as open source.

If Apache want to deliver P2 firmware that would be substantially more involved. It is easy to suspect that Apache will ship with P1 because there has been no overt evidence that Apache has been working with the singular P2 expert, Rick (N1GP). This would leave development of P2 firmware as an exercise for Rick and the rest of the community. Anyone who wants to develop firmware will need a G1 available to them.

Any support necessary for the P1 firmware should be nil, unless Apache makes some gross error in its design. P1 firmware has always been very stable because, as previously mentioned, P1 timing closure is relatively straightforward.

Reading the announcement it states that the G1 is compatible with a control program like Pihpsdr. Yet, the latest source code of Pihpsdr doesn't 'know' a G1.

It is similarly likely that the G1 is functionally identical to the 8000. If this is true, then having a G1 hardware option would merely be a convenience to avoid confusion, although perhaps it is necessary because of the different amplifier power output rating.

However, as you point out, it is extremely concerning that Apache communications with the firmware and software developer community about the G1 seems to be non-existent. I have sent email directly to Apache, copying the dev's, asking about this very thing, and received no replies.

Wouldn't the G1 deserve its own section on this forum, it having little commonality other than its name and brand?

Not at all. Again, it should be nearly identical to the 8000. It's merely an 8000 with a smaller amp, different FGPA, no front panel LCD/Arduino, and, quite unfortunately, no external support for audio line-level I/O (although it is undoubtedly supported internally as the CODEC IC will equally undoubtedly be the same). Remember that every 7000 variant, and the 8000, share nearly identical Orion MKxx SDR boards.

Because 99% of problems and questions associated with our radios are associated with the operation and configuration of the client software, I've arranged the forum to motivate organizing posts along those lines. This way the same questions can be concentrated together. It would be terribly confusing to have the same questions duplicated across multiple, hardware-model-specific sub-forums.

Similarly, their is very little functional difference between the all of the Apache hardware models, from the original -10 all the way to the latest -10E "revisited". Thus the idea to concentrate all hardware specific questions and discussions in the "Apache-specific Hardware Discussions" section.

The only exception is the G2, which is why it has its own, dedicated forum. The G2 is a special case because of the additional complexity of the internal Pi compute module, the "middleware" that runs on it, and the major differences in how the firmware works and is updated.

The G2 section is already quite complex with G2's many configuration options of stand alone vs external control, with/without control front, old and new control front, with and without CM5 upgrade.
And soon a v3 Saturn PCB upgrade being introduced. The G1 may be a source of confusion in a G2 dedicated section.

Agreed. I don't know what I was thinking when I moved this topic to the G1 sub-forum

Thanks for motivating these additional thoughts and discussion!

@wu2o
thanks for elaborating on the various aspects of the G1, like community support and where to discuss in the forum, the G1.

Had not expected your association of this G1 with the Anan 8000 (DLE I presume).
When I look up the specs I see for the 8000 two AD converters and diversity reception. So it must have dual input filtering.
The only design I found through Google that uses the Cyclone 10 LP FPGA is called Radioberry, but that uses 12 AD bits.

Just because two SDRs use the same FPGA, or FPGAs from the same family of FPGAs, has absolutely nothing to do with their architecture or capabilities. It is not an effective way to compare things.

I'm not sure it's something you are inclined to do, but you might consider reviewing the schematics that are available on this forum. Start with the Hermes (-10 & -100). After that would be the Angelia (-100D). After that the Orion (-200D). And, of course, the various Orion MKxx variants (7000's, 8000, Andromeda).

If you undertake such a review, you might be surprised at how similar they all are. This is not to say that each version did not improve in some way compared to the generations before it. But those improvements are relatively small, often subtle, and by no means reflect a major change in performance. While differences can certainly be measured in the lab, in the real world, with a real world signal and noise environment, there are often no measurable differences.

Board layouts (especially on the RF/PA boards) have changed to improve both Puresignal performance (via lower crosstalk) and phase noise performance. Better OCXO circuits have also improved phase noise and eased the use of external references. A very significant change was from 20dB of ADC preamp gain to 14dB. There are other small changes here and there. And FPGA part numbers have changed primarily based on cost and availability as opposed to performance. But, overall, the architecture, interfaces, and bill-of-materials ("BoM") is little changed from the original Hermes.

Therefore, it is a very good assumption that the G1 will follow the same pattern. The only reason I chose to compare it to the 8000, rather than the 7000, is because the rear panel looks very similar to the 8000. Thus the G1 PA/RF board(s) are likely a derivative of the 8000 design.

I mentioned the FPGA not because of performance but because its programming needs to be done and supported somewhere and has details that are very specific to the exact rest of the SDR hardware.
It apparently not being 100% identical to a SDR we already know, it makes me wonder who made the gateware, will support and maintain it, and where the source code can be found.
The open source aspect and the ability to contact the makers in case of problems, or ideas to do even better, is an important factor in the choice which SDR to buy, certainly given AL's history as hardware manufacturer but not necessarily software maker and maintainer.
.

Upon even more review of the G1 advert on the Apache website, it also seems that:

- This is a single ADC unit, so no diversity.

- There may not be any preselector, only the "broadcast band filter". This is not a big deal IMHO, as all previous preselectors have utilized third order filters, which is barely better than nothing. I just run my Apache stuff in "bypass" all the time because of this.

- Rather than using an 8000 derivative PA/RF board, it might be that this is a stripped down version of the old Rev. 24 board design.