I installed an aftermarket power supply in one of my HD1000's. It is a Morex brand supply with an external power brick and is a popular replacement supply on this forum (see pictures below of install).I did not install a fan at this time.

I also had another unit that has an OEM supply that I rebuilt, one of the items I addressed in the rebuild was replacement the electrolytic capacitors with better quality ones. I also moved the power resister away from the grove of capacitors so they would not get so hot and a couple other items were addressed too. So for this test, I am not really using an OEM supply per se.

TEST-1: After 2hrs at room temp (77F), sitting on the floor powered up, side by side, I measured the top of the cases and this is what I got.
Morex = 89 on the left (over the supply) and 82 on the right.
Rebuilt OEM = 92 on the left (over the supply) and 82 on the right.

TEST-2: I ran each unit for an hour in my entertainment system with the door closed and used SlimRoku to play music served from my PC, as well as watch a slideshow of pictures off my PC at the same time. I did this for each unit. I ran the OEM one first, then the Morex. The temp in the cabinet was ~80F.
Morex = 97 on the left (over the supply) and 89 on the right
Rebuilt OEM = 101 on the left (over the supply) and 91 on the right.

Today, I ran each unit in the cabinet for several hours and the data matched TEST-2 from last night.

I would really like to be able to do the same test, but with an untouched OEM setup.

A couple of thoughts...
1. The Morex didn't run that much cooler then the rebuilt OEM supply.
2. Does the OEM supply run a lot hotter before it dies then when it's new or rebuilt?
3. Are most people running a fan with their Morex supplies?
4. I need to get a hold of a working unit that has an untouched OEM supply in it and see what temp is runs at.

So in closing, what are other Morex users seeing?

What temp does your unit run with a Morex in it and do you use a fan?

Is your unit warm to the touch or cool? I would assume that cool would be 85F or less.

I did see a little better responce with the remote control with the Morex (like others have stated), but I also saw that (better range and snappier responce) with one rebuilt OEM unit that had it's 3.3V at 3.4V. Most of the OEM supplies are 3.25V (I want to say the Morex was 3.35V). So I wonder if the better remote range is more due to the fact that the voltage is at 3.3V, and not lower? I am pretty sure the 3.3v goes to the ATI ARM/CPU, I know the remotes IR codes do. I need to look at the OEM supply some more and see if there are any resistors that can be change to change the output voltage. There are no pots to adjust on these suppllies. I assume there is some sort of adj to the 8pin controller DIP that can control this, I just need to look into it some more.

I also know that if the OEM supply has enough noise on it, that the remote will stop working all together. So I wonder as the supplies get closer to failing, do they get noisier too and cause the remote's performace to go down hill too?

What I may do next, is take the OEM supply that has the 3.4V out and put it into a unit that has a slugish remote and see if it makes it more robust or not.

Regards, Matt

Here are the pictures of my Morex install in my HD1000. You can click on the link below each one for a larger view.


http://www.wizard-widgets.co.uk/roku/ma ... e-1024.jpg


http://www.wizard-widgets.co.uk/roku/ma ... t-1024.jpg


http://www.wizard-widgets.co.uk/roku/ma ... t-1024.jpg


http://www.wizard-widgets.co.uk/roku/ma ... r-1024.jpg


http://www.wizard-widgets.co.uk/roku/ma ... r-1024.jpg


http://www.wizard-widgets.co.uk/roku/ma ... SU-800.jpg


http://www.wizard-widgets.co.uk/roku/ma ... e-1024.jpg


http://www.wizard-widgets.co.uk/roku/ma ... p-1024.jpg


GOT ROKU?


http://www.wizard-widgets.co.uk/roku/ma ... k-1024.jpg


http://www.wizard-widgets.co.uk/roku/ma ... t-1024.jpg[/img]

Matt,

I can tell you that the OEM supply in my Rev B box ran a LOT hotter than the Morex supply. In my AV cabinet, I had to install a fan in the back of it just to keep the PB cool. If not, the top of the case was too hot to touch.

With the Morex supply, I can let the box run for hours doing similar to what you did and touch every component on the Morex supply w/o getting burned. That's with no additional cooling.

I found the output electrolytics on the OEM supply were basically desoldered. It got so hot that the joints failed!

I have no fan in the PB itself.

I'm not sure if you are measuring the temp on the top of the Roku case or at the supply itself. If you can rent or get hold of an IR thermal scanner, I think you might find the results surprisingly different. I am sure the OEM supply has many more hot spots.

Jon

I like what you're showing with the switch/input module. Could you give a bit of detail about it and what benefits it brings?

dgburns wrote:

I like what you're showing with the switch/input module. Could you give a bit of detail about it and what benefits it brings?

AFAIK, the switch just allows you to completely turn off the HD-1000. Some people like to do it for heat reasons when they aren't using it.

I'm speaking for Matt here, but I can't see any other benefit to a switch...

na9d wrote:

dgburns wrote:

I like what you're showing with the switch/input module. Could you give a bit of detail about it and what benefits it brings?

AFAIK, the switch just allows you to completely turn off the HD-1000. Some people like to do it for heat reasons when they aren't using it.

I'm speaking for Matt here, but I can't see any other benefit to a switch...

Thanks!
I don't know what benefit it really brings per se. I just did it because everyone else was doing it or asking for it (I wanted to compare what people were doing Morex wise to the OEM setup). Plus with the provisions on the mini ATX supply (PON), it made sense. I would of had to tie it to ground to get the Morex to turn on anyway.

I assume like na9d said, to turn it off for heat reasons. But the Morex doesn't seem to run hot and it's only a couple degree cooler then a rebuilt OEM setup. I really assume it's because we all like to turn something off when it's not in use, and if it were not for the 20-60sec boot-up time, I would agree.

I took a piece of anodized aluminum and made/drilled tabs in it to mount to the two OEM stand-offs that were on the bottom of the chassis (two of the four that held the OEM supply in place). Then I drilled a hole for the Morex plug and the Radio Shack switch (there an alignment notch too in the hole for the switch). There was an extra ground tab on the Morex plug and I used this to tie the Morex to the Chassis as well as a second wire from that mount for the ground for the PON wire (the green one). I used the insulation plastic from the OEM supply too. I had to drill two mount holes in the bottom of the chassis to hold stand offs for the Morex too. I ordered the same plug and pins from Digi-key and made my own harness and elected to use Orange and Red, verse the OEM Orange and Yellow since the yellow wire is 5v and not 12 like yellow normally is.

Also, as another fine forum member ( olllllo
) brought up, several reasons...
Reasons to install the toggle:
1) Because you can.
2) To recover from a hairy lock up.
3) Vacation Mode.
4) Kid Mode.
5) Can't reach the cord.
6) Save the earth from green house gas.
7) Xtra modding points
Fun to say, "I toggled my Roku this weekend," to friends.
9) Desire to give Radio Shack your zipcode and address.

let me know if you have any other questions, Matt

Oh yea, I also designed the Morex setup to be a drop-in for easy change out, verses having to drill holes in the chassis to hold/mount stuff. The hole assembly could be made and then put into any unit with only having to drill two holes on the bottom of the case for the stand offs. Even that can be done away with if I wanted too. I could mont the whole thing on a thin piece of plexi and use the OEM stand-offs for hole the plexi, that in turn held the supply and the bracket too. So it could be a kit per se. It was not a quick project though.

I understand and fully support (even having suggested for your original work with the other supply) adding a power switch. I don't understand the power in port though, it LOOKS like a DC input port but seems to be AC input? So there is a custom cable for this to go from wall to this input connector? Or is it really DC input, and there's an external brick/block?
Sorry I haven't followed the Morex threads enough to know if it's an AC-DC supply or just a DC-DC converter.

Doug,

The Morex takes 12 V DC input. There's an external AC to DC converter supply. This is the beauty of it in that the "hottest" device is outside of the HD-1000.

Jon

na9d wrote:

Doug,

The Morex takes 12 V DC input. There's an external AC to DC converter supply. This is the beauty of it in that the "hottest" device is outside of the HD-1000.

Jon

DG, Now I understand your questions!
Yes, as Jon said, there is a power brick that comes with it. It converts 120VAC to 12VDC and then plugs into the back of the HD1000 with the special plug that came with the Morex kit.

Jon, on my unit the power brick is not that hot. I would not even call it warm, the hottest portion of the whole Morex suppluy are the three chokes that are on the PCB. I'll pull the cover off weekend and use my IR temp gage to measure the three chokes and also measure the VR's on the Morex (I bet those are hotter then the power brick too) and then I'll measure the brick and post results. On both the Morex and the OEM supply, the AC-DC conversion does not make near as much heat as the DC-DC conversion.

Morex- Chokes and VR's (+12, -12, +3.3,+5 and -5volt).
OEM - Regualtor (+5V) on the heatsink, MOSFET on heatsink, rectifier on heatsink, DIP controller (+3.3V) and the power resistors on the output.

As I said, let me measure some temps this weekend and get back to y'all because I bet the inductor (black round core with think copper wire around it (just like the one on the OEM supply) gets hot too.

To further explain is may be true, the OEM supply has about 11watts of disipataion and the Morex is about 10watts. since the brick is not the hottest part, and it's the DC-DC portions, then that is probably why the heat numbers I got for both the OEM and Morex supply were only off by 3-4 degrees.

CAVEAT: This is still just "theory" unitll I get numbers this weekend.

Reagrds, Matt

I'll also try to retake the picture of the rear of the unit with the power brick pluged in and replace the current picture from the first set.

OK, the power brick was ~97-98F on all four sides (checked many areas in each plain).

There was a portion, about a quarter size on the bottom of the supply that was 100F.

On the supply (DC-DC portion), if you start at the top where the power (12V) comes in and work down to the ATX plug (180 from the picture above) this is what you get...
3 caps 97-99F
Inductor (metal core with wire around it) 118F
left and right VR's 115F anf 110F
IC in between 117F
left and right VR's below that 117F and 115F
then there are two round chokes, the left was 150F and the right was 170F, there are 4 caps in the middle (green majic marker on top) that wer 118F, 120F 118F and 110F
then a VR in the middle at 120F with an IC on each side at 133F and 121F
A sideways choke below that was 112F then an IC at 100F and then 3 caps at 100F

So the PCB of the Morex; which is in the HD1000 is between 97F-170F. The supply was measured and there is a 10watt draw on the power brick, whereas the OEM one has a 11watt draw from the wall.
Regards, Matt

Here is picture of the orinataion for the above.

I had noticed how hot the top left (above PS) was on my HD1000 so was not surprised when it failed - especially since the power supply is ALWAYS ON even if you think you've turned the unit off with the 'power' button.

After reading all these posts and getting my unit repaired I plugged it into a switched outlet on a:

Smart Strip Power Strip
With Autoswitching Technology

Seven outlets work together, autoswitching your devices on/off automatically, to save you money on your electric bills. 1225 joules of surge protection keep your equipment safe from even the harshest power spikes. Conveniently-spaced outlets and a 45-degree, angled space saver plug make the Smart Strip a friendly addition to your electronic systems.

and plugged my power amp into the control socket.

This way my HD1000 is always REALLY off when I'm not using my AV/home theater. Doesn't really solve the heat problem but should greatly prolong life as PS is not running 24x7!!

DOWNSIDE: if I turn on my AV and want to go straight to Roku, have to wait the 30-60 secs. for it to 'boot'...

I am very surprised at the results Matt is showing regarding heat. I can tell you that when my Rev B original supply still worked, the top of the PB case was so hot you could barely touch it unless I had air moving through my AV cabinet. With the Morex supply, it's always cool to the touch.

Let's look at efficiencies.

The PB is rated for an input of 100 to 240 V and 1A. Typical AC voltage in the USA is between 115 and 120 Volta. From basic electronics we know:

Power = Voltage * Current

So 120 * 1 = 120 Watts. The input to the PB took 120 Watts.

If I remember correctly, the PB stock supply lists the outputs as:

3.3 V at 7 A
5 V at 1 A

This then would equate to a total power consumption of: 3.3 * 7 + 5*1 = 28.1 Watts. Therefore the Roku power supply at max will dissipate almost 92 Watts of power! That's HEAT and heat inside the chassis.

Now, the morex supply.

The power brick of the Morex is rated at 100 V to 240 Volts AC input at 1.8 A. This would be:

120 * 1.8 = 216 Watts.

The output is rated at 60 Watts maximum. So under maximum load, it could dissipate 156 Watts. It would be quite warm, but it's OUTSIDE the chassis.

Now, for inside the chassis.

Assume then that if the output of the brick is 12 V at 5 Watts, that the input to the ATX supply is rated the same.

Now, our 3.3 Volt line is rated at 5 A = 16.5 Watts and the 5 volt line is rated at 1.5 A = 7.5 Watts. This equates to 24 Watts of power.

Therefore the INTERNAL dissipation of the supply is 60 - 24 = 36 Watts. This is a LOT less than 92 and is a lot less excess heat.

Now, keep in mind that there's several outputs on the ATX supply that aren't even used that all fall into that 60 Watt rating. So even at maximum draw, from the Roku, the input power is probably far less than 60 Watts.

Additionally, I think the layout of the Morex supply allows the components to dissipate the heat more efficiently and transfer the heat to the air better. The layout is more spread out and not so compact and it's not enclosed in a metal case like it is with the original supply.

This is why I think running w/o a fan using the Morex supply and leaving the supply on all the time is fine...

na9d wrote:

So 120 * 1 = 120 Watts. The input to the PB took 120 Watts.

Yes, BUT current is a "draw", whereas voltage is a "pressure". The voltage is for purpose of this discussion a constant, but the current is whatever is required. Just because it is spec'd to draw 1 AMP doesn't mean it ever does draw that much. So going in saying the OEM supply is 120W of potential heat is sorta misleading, and saying the OEM supply has an excess of 92W to dissipate as heat is an inaccurate assumption. If under load it's only drawing 1/2 an AMP from the wall, that'd be only 60W and dissipating 32W as heat in your example. We've seen evidence of this variance in many reports from Rev B owners such as yourself that your unit was "hot" to the touch when running, while others (such as myself) reported their Rev B was only modestly "warm" to the touch when running. I have no explanation for WHY this happens in what seemingly should be identical units. I'm sure if measurements of current consumption had been made those with "hot" units would have seen a higher current draw from the wall than those whose units were only "warm". All of which, in the end, seems not to matter since both "hot" units and "warm" units fail with seeming inevitability!

We'd need to measure actual power consumption under load (I'm sure someone has done so), which I would do if it weren't for my Rev B being dead.

I am very surprised at the results Matt is showing regarding heat. I can tell you that when my Rev B original supply still worked, the top of the PB case was so hot you could barely touch it unless I had air moving through my AV cabinet. With the Morex supply, it's always cool to the touch.

I am not doubting at all about how hot your unit was running. There seems to be a correlation between the OEM PSU stating to fail and it running hotter. Earlier in the post; I thought I brought that up, I’ll have to go back and re-read. Now the good news is, one of our forum members that has a working HD1000 but runs like a hibachi is going to send it to me to evaluate and after that, I am going to service the unit.

Let's look at efficiencies.

The PB is rated for an input of 100 to 240 V and 1A. Typical AC voltage in the USA is between 115 and 120 Volta. From basic electronics we know:

Power = Voltage * Current

So 120 * 1 = 120 Watts. The input to the PB took 120 Watts.

First off, you pull current, you don’t push it. The supply is rated at that, saying that it can support that much draw, not that it’s going to force feed that much to what is on the other end. I want to say that the OEM supply only drew 0.09A not 1.00A. Let me verify that and get back to you (I was pretty sure the OEM supply was 11watts and the Morex was 10watts (9.6Watts)

If I remember correctly, the PB stock supply lists the outputs as:

3.3 V at 7 A
5 V at 1 A

Yes, that’s right, it’s RATED @ 3.3V @ 7A and 5V @ 1A

This then would equate to a total power consumption of: 3.3 * 7 + 5*1 = 28.1 Watts. Therefore the Roku power supply at max will dissipate almost 92 Watts of power! That's HEAT and heat inside the chassis.

Food for thought, the Morex is only rated at 3.3V @ 5A, so if the HD1000 needed 3.3 @ 7A like you are implying, then it would not work with the Morex.

Now, the morex supply.

The power brick of the Morex is rated at 100 V to 240 Volts AC input at 1.8 A. This would be:

120 * 1.8 = 216 Watts.

I want to say the brick has a 0.08A draw on it, not a 1.80A draw.

The output is rated at 60 Watts maximum. So under maximum load, it could dissipate 156 Watts. It would be quite warm, but it's OUTSIDE the chassis.

Now, for inside the chassis.

Assume then that if the output of the brick is 12 V at 5 Watts, that the input to the ATX supply is rated the same.

Now, our 3.3 Volt line is rated at 5 A = 16.5 Watts and the 5 volt line is rated at 1.5 A = 7.5 Watts. This equates to 24 Watts of power.

Therefore the INTERNAL dissipation of the supply is 60 - 24 = 36 Watts. This is a LOT less than 92 and is a lot less excess heat.

Now, keep in mind that there's several outputs on the ATX supply that aren't even used that all fall into that 60 Watt rating. So even at maximum draw, from the Roku, the input power is probably far less than 60 Watts.

Additionally, I think the layout of the Morex supply allows the components to dissipate the heat more efficiently and transfer the heat to the air better. The layout is more spread out and not so compact and it's not enclosed in a metal case like it is with the original supply.

This is why I think running w/o a fan using the Morex supply and leaving the supply on all the time is fine...

I am just an Electronics Tech by trade (and education, I have a CET from Texas A&M (Yea, I know, there’s the problem )), let me contact Chris who is an Electronics Engineer and can probably explain this better, I may be causing more confusion.

I didn't see your reply untill after I typed miine. I have measured the consumption under load and it was 11watts for the OEM supply and it was measured at the AC input since the whole supply is in the chassis. The consumption of the Morex was 10watts and that was maesured that the 12V in to the Morex since that is the only porion that is on the Chassis.

As I said, let me double check and get back to everyone, my memory is not the best (and we will not even bring up my spelling )

dgburns wrote:

We'd need to measure actual power consumption under load (I'm sure someone has done so), which I would do if it weren't for my Rev B being dead.

Burkhardi wrote:

I didn't see your reply untill after I typed miine. I have measured the consumption under load and it was 11watts for the OEM supply and it was measured at the AC inout since the whole supply is in the chassis. The consumption of the Morex was 10watts and that was maesured that the 12V in to the Morex since that is the only porion that is on the Chassis.

As I said, let me double check and get back to everyone, my memory is not that best (and we will not even bring up my spelling )

dgburns wrote:

We'd need to measure actual power consumption under load (I'm sure someone has done so), which I would do if it weren't for my Rev B being dead.

If that's true, 1W of difference between the OEM and Morex isn't what's killing OEM supplies. However, I don't think anyone would argue that improving the heat dissipation capabilities of the individual components inside is not a valuable addition.

I wonder if anyone's looked at the DC output of the OEM on a scope? If it's the truly crappy supply seemingly implied by many posts on these forums I'd suspect we'd see a lot of ripple riding on the DC, which could very well be contributing to components (i.e. those reporting mainboards dying) downstream failing.