You need to provide details of
The cells used (Vmp, Imp, Wmp), how they are connected, and details of your load.
If there is a battery used, what are it's specifications (V nominal, Ah, chemistry) and how is it connected.
What controller are you using and what does it do?
What MPPT controller are you using and what will it do?
Assume a non-MPPT controller to start. The following can be modified slightly if MPPT is used - which is discussed at the end.
PV systems with a 12V LA (lead acid) battery usually have PV panels rated at Vmp ~= 17V-18V (with 18 PV cells in series as mkeith noted).
A 12V LA battery requires more than 12V available as it needs to be able to be charged to 13.X volts for float purposes and to 14.xV for boost charging. Add some drop in the controller and wiring and connections and 36 cells is what is often used.
If your load is NOT a battery and you can tolerate dropouts (which would not suit a computer system with no battery) then your panels can be rated at closer to 12V for eg a 12V inverter. A panel of say Vmp=14V will produce most of that voltage at lower solar insolation levels if only loaded to the level of energy available. (ie Vmp at < 100% solar insolation is close o Vmp at 100% insolation).
To "just achieve" 12V out you'd need about 24 PV cells in series at about 0.5V each loaded. As you say you have 100 cells that suggests you may have wired them as 25S4P to give Vmp of ABOUT 12.5V. That is too low for a 12V system and much too low for a 12V LA battery based system.
You could rewire 99 cells as 33S3P to give Vmp of about 16-17V, and Imp of 3/4 of what you have now.
When adding PV cells in a series string, the available string current will be that of the lowest Imp cell on the string, as Brian Drummond notes. This means that cells added in series MUST have the same Ip current as the existing cells (or higher)if existing current rating is to be maintained.
An MPPT controller takes the available PV energy and converts the input voltage and current to some other voltage and current while attempting to optimise the load on the panel to achieve maximum power output for a given solar insolation level. Energy out will be slightly less than Energy in due to inefficiencies in the MPPT controller. ie Vout x Iout < Vin x Iin. Efficiencies of 90% should be achievable.
In addition to controller losses a 12V to mains AC inverter will have losses. Efficiencies of 80*+ can be expected.
A 400W panel will produce 400W in full sun when optimally loaded, orientated to optimum position, clean, new and at 25 C. When not optimally aligned, not full sun, used, somewhat dirty and sunwarmed output will be lower and ma of course be much lower.
The cells are 0.5 volts 8 amps so 4 watts each. They are wired as 4p25s and rewiring is not an option as they are permanently fixed.
I have no battery or controller at this moment.
I would prefer not to use extra cells in 4p11s as the extra wattage will underutilize my 400w inverter.
Can the MPPT controller raise the voltage to 17v or so and prevent voltage drop? –
Capabilities of MPPT controllers vary but in this case a suitably flexible one will take the available Vin (of maybe 12V) and convert it to whatever voltage is needed for battery charging or powering an inverter.
Load, battery, inverter and MPPT controller specs need to be known to comment usefully on what can be achieved.
In your situation the MPPT controller will be a boost converter (as Andy suggested was needed). It need not boost to 17V (which relaates to meeting all aspects of lead-acid battery charging) but rather to whatever voltage is actually needed. Whether it can provide the input voltage you need under load depends on actual available energy and actual load. As noted above, any added cells need to be of at least the same current rating as the existing ones.
The panel should be oriented so as to achieve maximum insolation utilisation. This is usually achieved by 'pointing it' due geographic (not magnetic) South in the Northern hemisphere setting it's azimuth angle for the lattitude concerned so the panel points at the sun's position at true midday.
If you can track the sun across the day so much the better but this is not usually done. If azimuth angle can be varied with season this is also useful but less so.
Output at a given moment compared to Wnowmax is reduced by misorientation by a factor of (1- cos(error_angle)) and NOT by sin(error_angle) which is far larger. ie at 10 degrees off-sun pointing loss is (only) 1.5%
Note: You need to take the gems that you have provided in various comments and edit them into your question as people are generally unlikely to scour the comments for extra material, some points can easily be missed even by people who make the effort, and the aim of the site is to provide quality questions and answers for future use. Even if one thinks the aim is to solve problems now one needs to dance to the site owners tune :-).
More questions and relevant ramblings:
It seems that you do not have an overall 'plan' and are serially doing things towards achieving your goal without having an overall picture of how they fit together - OR - If this is NOT the case then there is much you have not told us about.
You say That:
You are planning to operate a (presumably) AC mains powered computer from a solar powered inverter but that you have no low voltage battery or (as yet) controller.
You have made a nominally 400 Watt PV panel (not a trivial task!) with Vmp = 12.5V.
You subsequently found that the voltage needed to be higher.
Presumably you think or know that your inverter will not operate from a
Vmp = 12V5 (or actual 11V or ...) panel.
You MAY be right, and may not, but we don't know why you think so, whether you actually have the inverter, whether you have tried to see what voltage it really needs.
You gave us your basic PV panel specs but we don't know if you have measured what it actually produces. Real Vmp MIGHT be enough higher to be useful. Or not. We do not what wattage it really makes in full sun when optimally pointed or how you are going to point it.
Crucially, we do not know if the computer will run or crash when panel voltage fluctuates with clouds/snow/rain AND night-time.
This suggests you may also be using mains AC and this then becomes a 'grid tied' inverter. If it is then it must be commercially built and certified for grid connection. In NZ where I live - and in many other countries the applicable standards require the actual PV panel to meet stupidly rigorous certifications if used with a grid tied inverter OR attached to a residence. That may not apply where you are or not be relevant (but it probably does and is).
Have you measured the voltage/power curve in full sun?
Is the inverter grid connected?
Is it MPPT?
Is it commercial and certified?
Can you provide specifications and/or links re inverter and anything else relevant.
How does all this work?