ArtemisSDR-v2.1.1 <> v2.1.6 RMS average power SUNSDR2DX

[ELSDR]

Hello

I would like to mention what I have noticed since version v.2.1.0 is ok onwards all version have lower RMS power around 10watts.

I did a test between ESDR2 and ESDR3 and Artemis SDR (older and newer version) on 40m 7100kHz Drive 100% Dual Tone Test freq 700Hz and 1900Hz.
Older versions below 2.1.0 have exactly the same RMS power on an external wattmeter as ESDR2 and ESDR3 on Drive 100.

I know that the version v2.1.1. and up has a reduced gain >> Power calibration: voice MOX wire amplitude is ~1.4 dB lower than v2.1.0 (the IqGain headroom drop). You may need to push the drive slider up by ~17% on voice to get the same on-air watts you had in v2.1.0. TUNE behavior is unchanged — the cal anchors remained locked to the 1,167 gain.

Push the drive slider up by ~17% on voice to get the same on-air watts you had in v2.1.0 hmm the slider is already at 100%... ? only the radio is in use, no amplifier ;)
PA settings profile can only be changed downwards and not upwards example from 37.7 can change the number, but the power does not change.

Yes, the TX meter in ASDR from version 2.1.1 shows 50w with dual tone test, which was not the case on older versions of the TX meter, but from this version the RMS power on the external wattmeter is different-lower.

Additionally the tune slider on 100 tx meter shows 70w on newer version for comparison in version 2.0.16 around 100w which is okay value.

Additional question for everyone else.. ;)
Does anyone use the sound card from RME and the ASIO driver for ArtemisSDR ?

ASIO works for me in ESDR2 and ESDR 3 but it doesn't work for me in cmASIO.
I activate ASIO device in Audio settings cmASIO, restart ASDR and only hear a quiet distorted sound that is completely unintelligible.

SUNSDR2DX
fw 88.8
ASDR current version 2.0.16 higher avg power
Win10
RME Babyface PRO FS last fw 322 last drivers 1.276

73 Elvis

[kk68]

Hi Elvis, thanks for the detailed side-by-side against EESDR2 / EESDR3 — that comparison is genuinely useful. Let me put the v2.1.1 change in context, because what you're seeing is the intended behavior, not a regression we missed.

The SunSDR2 DX (and PRO) hardware doesn't expose a TX feedback path the way some other SDR transceivers do. Without a feedback sample of what the radio actually puts on the air, adaptive predistortion — the proper way to clean up TX spurs — isn't possible. This isn't an Artemis limitation; it's a hardware one. EESDR2 and EESDR3 run on the exact same hardware and show the same spur profile because the underlying chain is the same. Expert's own documentation acknowledges the radio gets noisy in close-in IMD above roughly 40 W of output. That's the radio.

What v2.1.1 did was bring the wire-encoder side of the chain to its cleanest achievable state given that constraint. The voice IqGain dropped from 1.167 to 1.0 because the higher gain was hard-clipping the wire-side quantizer on multi-tone speech content and generating IMD3 / IMD5 spurs that weren't even coming from the radio's PA — they were artifacts we were creating ourselves before the signal ever left the PC. Dropping to 1.0 eliminated that source of spurs. The cost is a ~1.4 dB lower wire amplitude on voice, which translates to slightly less on-air watts at the same drive setting. For an operator running barefoot at drive 100%, that's the 10 W gap you're seeing.

There is an immediate workaround for barefoot operators who need the watts back. Go to Setup → PA Settings → PA Gain (By Band) and bump the dB value for the band you operate on by ~1.4 dB. That scales the drive curve and effectively gives you back the 1.4 dB the IqGain change took out, without re-introducing the close-in spurs (because the wire-encoder side stays clean — the extra gain happens downstream of the quantizer). Test in TUNE first against your external wattmeter, then verify on voice. If you want to be more conservative start with +1.0 dB and walk up from there.

For the longer term, I've been working on a different path: an external TX sample tap (the LP-500 sense-port retrofit, ~$75 of hardware) that would give Artemis a real TX feedback sample. With that in place, adaptive predistortion becomes possible and the headroom can come back natively, with much better spur performance than v2.0.x. Hardware is on its way; I'll have more to say once it's in the lab.

Until then, the v2.1.1 design stays as the default. The trade-off — modest output reduction for measurably cleaner close-in spectrum so we're not QRMing neighbors on a noisy radio — is the right one for the community as a whole. If individual operators prefer to maximize output at the cost of the spurs, v2.0.16 remains available, EESDR2 / EESDR3 are valid options on the same hardware, and the PA Gain (By Band) workaround above lets you keep v2.1.x and the clean wire encoder while recovering the watts.

On the separate TUNE-power observation: the 70 W you see on TUNE at drive 100 in v2.1.x vs 100 W in v2.0.16 is actually the v2.1.0 per-band TX meter cal correction, not a real power change. The earlier single-curve meter cal was over-reading by up to 28 W on 40 m. The radio is putting out approximately the same actual watts on TUNE; the meter just stopped lying. Easy way to verify: compare external wattmeter / EESDR3 / Artemis on TUNE at one drive level on 40 m. If the external wattmeter and EESDR3 agree at ~70 W and Artemis also shows ~70 W, that's the cal working correctly. If they disagree, that's a separate cal issue and I want to know.

For the PA settings profile slider that only goes down — I need a screenshot or the exact menu path to find which slider you mean. Setup → PA Settings → PA Gain? Setup → Tests → PA Cal? Somewhere on the front panel? A screenshot of the panel with the slider visible would let me pinpoint it. If you were looking at Setup → PA Settings → PA Gain (By Band), it accepts negative offsets — so pushing it down should work — but it should also accept positive values; if it doesn't, that's the bug.

For the cmASIO + RME Babyface PRO FS issue — a few diagnostic questions: (a) what sample rate is your RME currently set to in TotalMix / RME settings (48k / 96k / 192k)? (b) what cmASIO buffer size does Artemis show in Setup → Audio → cmASIO once it's active? (c) does switching the RME hardware rate change the distortion at all? (d) is the distortion present immediately on Artemis startup, or does it develop after RX has been running for a while? cmASIO runs at the SunSDR's native 312.5 kHz wire rate and resamples to whatever rate your RME is set to — most cmASIO issues I've seen have been at the resampling boundary, so the answers above should help narrow it down.

Radio firmware 88.8 — noted, thanks.

73, Kosta