NOTE: I'm using the 12 years per mod now, and also I'm a dreamer about the length of modding that could be applied, so some crazy arguments could come out from this:
I have been thinking about this... all samurai should be recruited in one two turns tops, because they are the land lords that have been training all their lives in the use of weapons. They are "ever ready to fight" guys, only waiting to be called upon service for their overlords, and once recruited they should have some experience right of the bat, not that much but some, and even some extra according to the dojos present on the region they come from. And for the weapons and armor, as they buy it for themselves it should depend on the wealth of the land owner class on the region. A wealthy class in the region means that the samurais (the land owners) are rich and can buy expensive armor and weapons. A poor region should have its samurais with (relative) low quality armor and weapons.
On the other hand the ashigaru should need more turns to be conscripted, since they have to be gathered, and then trained (assuming they have not served before, they where just peasants before recruitment). So a mob of peasants will be ready in one turn. Yari ashigaru, being the most basic weapon should need one month or two, while yumi would need more months of practice, at least 4 I think (unless they where bow hunters), the teppo is even easier so two or one month would do it too. The speed of this can also be influenced by the population numbers on the region. The officers in provinces with excess of population should have no problem finding young people to fill in the ranks, maybe lessing the training time a month, while a scarce populated area will add a month or more to the training. Their experience should always be green as they really are green, having never used a weapon of war before. Their armor and weapons should be relative to the local weapon and armor smith (built by the player) since this is the "government one", that provides armor and weaponry to the conscripted troops, also the stats like melee damage and charging should be increased by the encampment building and it's upgrades, but this obviously would also add some months of training.
So if you want really light and cheap ashigaru, don't make either armory or encampments/barracks/etc on the province and recruit them in a really populated one, you will be filling stacks really soon, they will be fragile but faster to replenish too, so if you have the population and the will to do so, what is stooping you? Or the other way around, you could build professional ashigaru armies by making barracks and armories on the provinces you choose to, but they will be more expensive and slow to muster.
All of this should also reflect on the units you see on the battlefield. Every unit should have a few models, like unarmored poor ashigaru, armored poor ashigaru, armored middle class ashigaru, unarmored middle class ashigaru, poor samurai, middle class samurai, rich samurai. And according to the status of the unit more models of one kind should appear in it on the battlefield. I.e. a samurai unit of 100 man recruited from a poor province would have like 80 poor samurai models, 15 middle samurai and only 5 rich samurai. This would be only aesthetically since the unit stats would be given at the moment of recruitment, but it can help to differentiate and give even more variety to the units on the battlefield. Maybe as they progress in experience they would be able to buy/scavenge better armor and weapons, so they start to look better in the battlefield and have better stats? I don't know, I'm still dreaming.
And not only the ashigaru would be slower to recruit in this system, but also more expensive, since you are the one paying for their equipment and upkeep, while the samurai would pay for most of their stuff and then only will require the upkeep while being mobilized.
This is pretty unbalanced and crazy right? A upside down kinda of world? Well even being faster and cheaper to recruit than the ashigaru, the samurai would not be as viable as the ashigaru, at least in the early game. See, all this is based on the fact that (two types or more of) populations can be added to provinces and then affected by recruitment, so their Achilles' heel would be the fact that you will never be able to mobilize as many samurai as ashigaru.
You wont be able, let's say recruit 12K samurai's in a year and start seeding the terror on your enemies, since this unbalance on the population classes would lead to a revolt, unrest and less income from your provinces (no one is there controlling the peasants). This will cap the quantity of samurai on the battlefields with some true logic and not some arbitrary head count number.
In the end if you perceive that the immediate threat if gone, or that the troops at hand are not needed anymore then you disband them and they numbers go back to a town. Specially good if you are short of labor on the fields or nobles to run and control said fields.
As you progress over the sengoku era and more "tech advances" are made, you can research something like higher mobilization, or laws/edicts(tech) that blur the line between ashigaru and samurai (like it happened in real life), so samurai units would require less land owner population and more common population (common folks with weapon skills). Keeping this trend (if one choose to) by the late game most armies would be composed of "samurai" units, only that this samurai units will not require so much land owner population, having a lot of the ashigaru filling in the ranks like the lower kind of samurai, so we can see a transition from the levy feudal peasant armies to more professional full time armies as it happen in the real Japan.