[benj111]

>To clarify some of the discussion, it is not a counterflow heat exchanger, it is a regenerative type.

Can someone expand of this?

intuition tells me that a regenerative design can be no better than 50% efficient, and would be worse at recovering latent heat

[ljosa]

I think it means that it's like [Lunos](https://www.lunos.de/en/for-heat-recovery). The unit alternates between exhaust and intake every couple of minutes. The air being exhausted heats up a core, which in the next cycle warms the air from the outside. Lunos e2 is advertised to recover 90% of heat and 20–30% of humidity.

[benj111]

yes i get that, but say you have an indoor temp of 30c and an outdoor temp of 0c. the average of this heat exchanger is going to be 15c. so on average youre only cooling the exhaust down to 15, and heating up the intake to 15c.

a counter flow heat exchanger can get the temperature higher than the average because 30c exhaust is meeting partially warmed intake, and 0c intake is meeting partially cooled exhaust.

Unless theres a phase change???

[ilyagr]

Perhaps https://news.ycombinator.com/item?id=42428537 or its parent will make sense to you: connecting 3 or more perfect constant-temperature regenerative heat exchangers in a series would make the ones at the ends work at a higher/lower temperature (of the constant-temperature heat sink for those exchangers) than the one in the middle, increasing efficiency of the overall system.

I'm not proposing this as a practical design, but it convinces me that 50% efficiency is not the limit.

[benj111]

Maybe? but doesn't that just become a poor mans counter flow heat exchanger?

[ilyagr]

It's closer to a counter-flow heat exchanger, but you could still have air only go in one direction at a time. Say your indoor temperature is 20 degrees and the 3 heat exchangers are at 25, 30, and 35 degrees, and outside is 40 degrees (I'm thinking in Celsius, though 40 is a bit extreme). You blow air out until they cool to 20, 25, and 30 degrees. Then you blow air the other way until they heat back up to 25, 30, and 35 degrees, with the air coming in being somewhere in the 20-25 degree range instead of 40 degrees that the outside air is at.

This assumes that the heat exchanger has just enough thermal capacity so that raising/lowering its temperature by 5 degrees would get the air to the same temperature. In fact, it might be easier to imagine if the air doesn't blow continuously. Each chamber could fill up with air and wait for the heat exchanger and the air to get to the same temperature, before moving the air to the next chamber.

[bittercynic]

Though the average temp of the overall core may be 15c, there may be a thermal gradient along the length, so maybe the inside end averages 28c and the outside end averages 2c, or something like that.

[benj111]

Perhaps.

I was trying to reason it out, you have the thermal mass of the heat exchanger material, that you aren't going to be able to tune to every temp differential.

at the start of the cycle, it may be very efficient, but at the end of the cycle where the heat exchanger temperature is going to approximate the temperature of the air, efficiency is going to drop, and that would be happening along the length of the heat exchanger, so by the end of the cycle the inlet would be basically doing nothing, and the exhaust quickly getting to that point (in the best case).

I was starting to confuse myself at this point, so the constant temp model seemed easier to reason about.

Although, having written all this the model would basically be a constantly shortening counter flow heat exchanger? with a thermal mass not necessarily tuned to the temp diff.

So its still going to be worse than a counterflow heat exchanger, at least theoretically.