Nick,
I won't repeat my comments from the other forum, since you seemed to get the gist of my explanation and understand the full-power tuning process. At that point it wasn't clear that you wanted to run reduced power, so let me tackle that now. First of all, can you say why you want to run lower power? For example, is it to satisfy a licence condition or are you trying to drive a linear amplifier and, if so, is the linear a valve (tube) type or a solid state type? The reason I ask is because there are several ways of achieving the reduced output; the best one is a bit application dependent. Also, what mode(s) do you wish to use - SSB alone or SSB and CW?
The most obvious way to lower the output is to tune for maximum output, then lower the drive to the power amplifier stage, in your case 2x6146B valves. For CW you do that via the "drive" control, which controls carrier levels when there is a carrier. For SSB without the processor, you use the mic gain to make the speech peaks reach the desired level, noting that average-reading power and current meters will read less than half (typically a third) the actual peak levels on uncompressed speech. Your FT101Z also has a nice RF speech processor and, on quick scan of the manual, you set the input levels to that using the mic gain and the output levels with the "drive" control. If you use the processor, make sure you read the manual and set up the ALC etc properly.
There is no escaping the reality of the valve dissipation calculated by Hamhock75 (above) but if you operate carefully with SSB, using uncompressed or very moderately compressed speech, you can keep the plate dissipation of each 6146B at the rated 35W for this class of service. Some radios intended for certain classes of Japanese operators often included a tapped HT transformer so the HT could be reduced for lower power operation within tube limits but I'm not sure if the 101Z has such an option.
Another way of controlling the power output is to use the ALC line on the "accessory" plug on the rear panel. Normally, as an external linear amplifier like the FL2100Z approaches full output it generates negative control voltage which is fed back to the FT101Z exciter via the ALC line, limiting the drive to the amplifier. You can use the ALC line in conjunction with a 9V battery and a pot to set the FT101Z power output. Note that I have never used a 101Z but I did use this trick with my Kenwood TS-830S, which is a similar hybrid rig.
Finally, if you are doing something more critical, like driving a solid state linear amplifier, the most foolproof way of reducing the exciter output is to run the radio at full power, then reduce the output via a 3 dB power attenuator (50 or 100W), something you can build or get off eBay for $80 or so. With an HF transceiver you can install the attenuator at the antenna connector since it has minimal effect on HF receive except for a small reduction in signal-to-noise ratio on bands above 14 MHz if you happen to live in a very RF-quiet location. (You can of course come up with a T/R switching scheme for the attenuator if you want).
I used an attenuator between my TS-590S and KPA1500 ampliifer but, since upgrading to a TS-890S with a very good and reliable power setting and limiting implementation, I've taken the attenuator out of service.
In trying to keep things simple I've glossed over a few deeper points in the explanations above but, in my judgement, it's close enough for what you're seeking.
Finally, if you have not yet found it, I'd have a look at the Fox Tango International web site (
http://foxtango.org/ft101zd/FT-101ZD.htm,
http://www.foxtango.org/ft-library/FT-Library/FT101ZD-901-902/FT-101ZDSG3.pdf, and associated pages) which has lots of good material on the FT-101Z. In particular, you should check that a "death" capacitor coupling the RF drive to the PA grids has been replaced. Apparently the original capacitors broke down, applying HT to the grids, destroying the final and, often, power transformers. As I mentioned, I've never had a 101Z but that sounds to me like a "must do" modification. In a somewhat amusing twist the notes outlining the 101Z idiosyncrasies also contain a nice section on how to seek help for faults (etc.) via internet forums.
73, Peter.