In the fall of 2019, another robotic milking system was started up, this one on a small farm in western Pennsylvania. Here are some notes on the initial transition from a conventional herringbone milking parlor to a single robot retrofitted into a freestall barn originally built in 1996.
Construction to accommodate the new milking system began in the spring. Around the same time, an automated alley scraper was installed to clean alongside sand-bedded freestalls and to help minimize cow disturbance, since the cows are no longer making group trips to the milking parlor. Delivery of the new Lely Astronaut A5 unit was in late July, allowing two and a half months for installation.
Over the month before startup, the milking herd size was intentionally dropped. Some cows were culled; others were dried off the day before startup to bring the number of milking cows to exactly 60.
Transition day began with a final milking in the now-retired milking parlor, finishing at 7:00 in the morning. The enthusiastic first cow entered the robot just a few hours later, at 10:30.
For the next couple of days, there was a crew working with the robot and cows around the clock. On the cow side of things, there were two to three people moving cows into the robot for the first 12 hours, which is how long it took to complete the first milking of all 60 cows. In hindsight, a two-man crew was adequate for working with the cows. The owner, who spent most of the time on the equipment side of things, was available to lend an additional set of hands when difficult cow issues arose. A crew of robot-familiar professionals was also on hand for the first few days to work with the owner on the equipment side.
Throughout the first milking, some cows spent quite a bit of time in the box. The robot frequently struggled to find teats and attach, especially with rear teats close together. With smaller cows, there was often a lot of forward and backwards movement, making the attachment and milking process even more difficult. Several cows were not comfortable with the robotic arm moving underneath of them and fought it. The arm took a beating but was impressively resilient. There were several times when a cow would manage to step over the arm with a hind leg, inhibiting progress until she stepped back over.
Amidst the frequent kicking, there were times when the clear shield in front of the teat detection laser became dirty, causing further delays in getting the unit attached. Reverse tilt was the other difficult issue for the robot to figure out, because the laser can’t detect the high-hanging rear teats when shone across the floor of the foreudder.
That first day, about one third of the cows did not let their milk down well. Of course, some of them were coming in with just a few hours of milk to begin with, but stress definitely had an impact, which was quite apparent in the robot records and the bulk tank.
Unexpectedly, during the first milking, milk from 2 of the 60 milking cows was not sent to the bulk tank by the robot. For one cow it was because she had been recorded as a fresh cow, and it was treating her according to the default setting, which is to separate milk for four days. It was never clear why it happened with the other cow, but it has not been a recurring issue.
The second time through the herd only took 9.5 hours. There was still a need for constant attention getting the cows to the commitment pen and into the box for milking, but a lot less hard pushing.
On Day 2, it was possible for just one person to manage cow flow most of the time, and the robot was much more efficient finding teats. One of the issues that came up was a bottleneck in the area where cows exit from the robot. Cows were congregating in the limited space, pretending to drink from the water trough, but really just curious to monitor activity in the robot area. This made it difficult for cows to exit at times.
On Day 3, an alarm went off for water flow and pressure at the robot. It was determined that the water pressure was less than adequate when water was also being used elsewhere in the barn, such as at the waterers. This led to plans for adjusting how water is stored and used around the dairy so the robot has a reliable supply at all times.
By Day 4, things were running well. Cows didn’t need excessive handling, they were getting used to the finger gates around the milking area, and the owner was starting to get more sleep. Some cows were so thrilled with their new setup that they would return to the robot more often than they should, checking to see if it would give them more grain. That night, the cows were left on their own for 4 hours, and 6 of them voluntarily went to be milked. The next night, the cows were left for 5 hours, and 16 voluntarily milked. The next night, the cows were left for 6 hours, and the robot reported that 22 had been milked during that time.
Two weeks after startup, the owner had settled into a routine of fetching about eight cows in the morning and eight in the evening, with some regular offenders. Getting fresh cows started was deemed a two-person job. Issues with bottlenecking or circling around the milking area were now only minor. Dirt problems on the robot’s laser seemed to be a thing of the past. And daily milk production was remarkably consistent from day to day.
Checking in again two months after startup, there were still a few fetch cows every day, but the process of checking reports and going after the important cows was easy to manage. The cows with the fewest average milkings per day (~1.4) tended to be the ones on that fetch list. Other cows were getting to the robot as much as 3.7 times per day.
According to DHIA data for the herd, somatic cell counts for several cows increased slightly on the first test after transitioning but dropped back down by the next test a few weeks later. As for milk production, the older cows tended to have a tougher transition, visiting the robot less frequently and having a more noticeable drop in production as compared to the younger cows. However, across the herd, milk production stayed fairly consistent.
There will surely be more observations and adjustments, but the new way of managing milk collection is off to a good start in this herd.