It's hard to say without any error messages, could be anything.
I'll try to give write up a test procedure to show you.
Okay, so based on your 40 block PoG difficulty window, there's opportunity to tithe several times a day.
Now, let's say some jerk wants to tithe 100 coins manually. This is more than the 75 txes you mentioned can go in a block.
Who decides which tx are included in the block? Does the miner decide? Is there a more equitable way to prioritize infrequent tithers first? Or maybe limit how many the same address can tithe in the same block and queue for the next block? This would increase illegal tithes because tithe too high or coin age requirement errors can appear. But it doesn't seem fair someone can tithe 100 times and block out other participants. I ask now because if PoDC is removed, all of that BBP will either move to MNs, PoG, or sold. So, those that participate in PoG means there will be more tithe tx in blocks overall.
22:35:56 tithe 1
22:35:57 
{
"Error": "This block is currently full of tithes. Please wait until the next block."
}
Yeah, Ill try to look at Capulos issue.
So wait; first of all I dont consider someone who tithes more than once a jerk. The system is deliberately set up to allow legal tithes in. The idea is for us to grow substantially, so that you really only have one legal tithe per day - then maybe per week if we had 20,000 users, etc. Its OK if someone has enough coin age to tithe more than once per block or day.
Your bringing up a plethora of issues all over the spectrum that are really bitcoin issues. They really are irrelevant. The bitcoin miner works fine in selecting the first 75 who make it in one block. The important thing to know is that if someone had legal coin age and a legal tithe, they have an equal chance of being in a block. (Because dash already went through this once, and they stripped the timestamps out of the miner, so the miners dont know who is first or later, miner only picks non-determistically based on current height and internal receive time - not on the transaction timestamped time).
I don't think you take into consideration what happens when we have more users; the diff rises and the frequency of tithing per user drops.