Power supply can cause slow Homey Pro 2023

I have tested three different adapters.

Soon I will test with 33W adapter, just purchasing it, if my HP2023 will boot up without giving me timeouts everywhere, I will report back my experience …

1 Like

Sorry; I mean restarting and/or power off/on to see if it flips back the switch

Do you guys think that a power supply that follows the Power Delivery or Quickcharge specification (with at least 15W) should work too?

Do they deliver a “normal” 15W when a non PD/QC device is connected?

I know, but as I haven’t found a power supply that can deliver sufficient power yet, I can not tell of the value is reset or not. I can’t tell if it didn’t reset the value or that the underclocking happens right away after restarting.

1 Like

Argh, just received the e-mail that my HP23 that was supposed to ship today but are postponed three weeks because of the powersupply issue. I’ve enough powersupplies to my other RPI4’s to power the HP23 with sufficient power, so this is not especially welcome for me

1 Like

Found one!
My iPad charger did flip (both) switch(es) back to “ false”.
Still CPU clock fluctuates a lot, so that charger is not really a good option.
But now we know these values are reset at a power cycle.

2 Likes

How long do we have to wait now. I find it frustrating you already prolonged youre test period. But no problem for homey the money is in the pocket.

1 Like

I bought the HP2023 because of the supposedly improved Zigbee performance.
As is turns out so far, Zigbee performance is even worse than with my HP2018.
Could this be related to the power adapter too?
I replaced it (with the result below) but I still have flaky Zigbee devices.

Never noticed a problem. Was using another PSU than the USB-C that came in the box. My previous model was the HP2016 so for me it was a big change in speed already. My HP2023 is very stable also the zigbee part…But will test with another PSU also.

2023-03-23 19_44_30-Homey Developer Tools

Contact the Athom support and ask :slight_smile:

Yeah, I’ll try to send them a request. Not gonna kill me if I do :wink:

@Henk_Renting so I guess it’s obvious… with new adapter, 1.5GHz, with old one…0.6GHz…

Power consumption wise it’s however not so significant…peaks are for 33W USB-C about 1A max…with original adapter I quickly measured about 0.7A max… (but really measured quickly, 30% less performance - yet, 1A is not 15W, it’s 5W - so I wonder why that original adapter is not sufficient, probably it’s caused by the power peaks)

Some readings with the new 33W adapter :
image
image

…and also fresh restore , no timeouts and all values FALSE:

image

I tried with a 100W PD supply and it still throttles.
I suspect the PD supplies won’t deliver the 5.2V that Homey is expecting - AFAIK it’s not part of the spec.

The CM4 is very picky, and there’s some voltage drop over the cable and pcb.

That`s what Emilie said on the developer hp2023 slack. So it’s has to be a power supply that is really good.

I just ordered the original Raspberry power supply, as suggested by Emile.
I’ll keep you posted!

2 Likes

Took a RPi supply and the clock vent up instantly and I don’t get the UnderVoltageOccured. So it will probably fix it for you as well.

2 Likes

I just wonder, is it VOLTAGE (Volts) or CURRENT (Amperes) issue ?
Just becuase it says “videoCoreUndervoltageOccured”…mail speaks about “voltage” as well…
Because if it’s about voltage, it will be not resolved so easily I guess.

The PCB and Cable “steal” about 0.1 volts, which is enough to cause an underclocked condition, so the problem is a voltage one.

Hm, I just measured it and it shows always 5.22V with the new adapter - regardless of the cable, LAN adapter or directly connected to the power adapter (didn’t measure it yet under load) - unfortunately while I’m trying to restore my ZigBee, I start seeing RED RINGs…

1 Like