The ANAN hardware using the Thetis software does a remarkable job with the AM mode and is capable of producing a broadcast quality signal using the built-in CFC audio software processing, if that kind of quality is what is desired. You can also make it sound like a stock Johnson Valiant too if that gives you a kick. My simplest understanding of what's going on is that Thetis produces a double-sideband signal and adds the carrier in at the end. So, unlike plate modulation, you are not actually modulating the carrier.
If you are using a PA power profile that produces a 100 watt carrier in Tune mode when the TX drive slider is set to the maximum setting, when you run AM with the ANAN set for maximum output the default AM Carrier Level setting of 100 will produce around 25 watts of carrier and your positive peaks will be allowed to reach just before the point where the DAC would clip, which will be right around 100 watts or slightly more. The PA in the ANAN has enough headroom to handle that signal and will be happy to do it all day along. It is worth noting that the algorithms that are in place in Thetis will absolutely not allow you to exceed 100% modulation in the negative direction with an AM carrier level setting of 100. It behaves exactly like a perfect "soft" negative peak limiter.
In the Transmit tab of Thetis, if you reduce the AM carrier level setting below 100 the carrier output will drop, but you are not affecting the sideband power that is produced, so the AM signal you are generating will not be any louder to stations you are talking with. Also, the further you drop the AM carrier level setting, the less that perfect negative peak limiting algorithm will exert itself to control negative peaks.
At some point, with an AM carrier level setting of less than 100, depending on how much audio you are using your signal will begin to exceed 100% in the negative direction. You will not splatter, but when observed with a scope you will see little wavelets passing through the baseline between your positive peaks. They will not be heard by someone receiving with a synchronous detector (SAM in Thetis) but if a diode detector is used, you will begin to sound distorted. If you go to extremes with a low AM carrier level setting, you will begin to sound like a quacking sideband signal to someone using a diode detector.
The short story with the AM carrier level setting is that you won't go wrong with the default setting of 100, and if your signal is observed with a scope, you will easily reach 100% positive peaks with 99% negative peaks that do not hit the baseline. If you use the soft-limiting processing in the CFC tools you can attain incredible audio density and loudness that most plate modulated rigs will not be able to produce without expensive audio processing hardware. If you want to generate a signal that "looks" like it has asymmetry I have found that an AM carrier level setting of 90 will make it look like I am modulating around 120% positive and there is enough resistance in the negative direction that I am not exceeding 100% in the negative direction. It looks like a broadcast station on a scope, but it's actually "fake" asymmetry because you are not gaining any additional power in the sidebands.
Note that you can certainly run pure signal while running the AM mode and it offers a pretty sweet bonus due to the noticeable improvement in THD. In my opinion, a major disadvantage to using pure signal with AM is that if you use the MON function in Thetis to hear your own audio you will only hear yourself in one side of your headphones and what you hear is the pre-distorted signal which is not what is heard by stations receiving you.
Hope that helps!
73, Rob W1AEX
NOTE: The YouTube video at the link below shows my station scope as my ANAN 200D is running AM with a wide range of AM carrier level settings.
https://youtu.be/069_uUSz3X4