PSU Dissected for Repair

[grandefuturo]

Update:

The display unit has a voltage regulator of the CM2596G family (3.3V or 5V).
There's also a 1.0F, 5.5V CAP on board (lifetime according to spec: 1000Hrs at 70C !!!!). Overall, the quality of the motherboard (made in Taiwan) looks much better thant the PSU/bass cabinet. Most part of the space is used by two mini stereo speaker plastic boxes, located on the left and right behind the front panel.

[grandefuturo]

ok, last message for today:

The CM2596 is a step-down switching regulator, I can't see whether it the 3.3 or 5V option. I suppose it's the 5V.

Input voltage is 7-40V, 45V max. That means it won't be damaged with the primary PSU failures mentioned here in this thread, except an input cap prior to the 2596 which is possibly a low voltage cap only.

Next I will open the display unit again, just to check all components prior to the 2596 regarding their voltage tolerance. Then I will perform a test with 19V only (for audio and microcontroller), checking the temperature of the 2596. If all test are successful, I will throw out the built-in PSU and go with a notebook PSU or another of-the-shelf PSU which probably fits into the bass cabinet...bingo.

[williMeier]

can you pls post the wire colour / voltage assignment

[grandefuturo]

Voltages are printed on the PSU board close to the cables. There's another connector when the cable enters the display unit, but the colors I give you here are for the cables soldered on the PSU board:


blue: 19V
black: GND
red: 9.5V (mine actually has 10.2V)
black: GND
white: shield /GND

all GNDs and shield have one common connection within the display unit, so there's no need to separate them.

[grandefuturo]

PSU DisplayConnector

19V blue red
GND black black
9.5V red orange
GND black(grey?) black
Shield white black

[S80_UK]

Thanks for sharing your findings. This is interesting stuff....

The difference from 0.54A to 0.37A corresponds to a permanent overload of almost 50%. This is definitely a system design bug, not a manufacturing issue.

Next I will open the display unit to check why it consumes constantly more than 0.5A, even if in standby. This bothers me the most.

In my view the current consumed is quite reasonable for a vacuum fluorescent display of that size - so the current draw does not bother me at all. A quick look at Itron data sheets confirms this. Much of the current goes in the filament for the display (there are quite a few filament wires in front of the display - glowing very dull orange if you turn all the lights off with the display dim but not off).

Having said that, though, a major concern would be the under-rating of that supply. We know already that the capacitors are highly stressed. If the current capability is much lower than needed, that only makes things worse.

Your observation regarding the 1.0F capacitor is not that critical, since it is what is known as a super-capacitor. Very different from a conventional electrolytic, and it would be used for keeping data in volatile memory, or for keeping a real-time clock running in the absence of mains power. It will not affect normal operation. And unless it is heated, it should be down at 30 - 40 deg C where it will have a massively improved lifespan.

Hope this helps.

BTW - I like the idea of using a lap-top supply. You will need to get from 19 - ish volts down to 9 or so, but it's an attractive solution.

[Burkhardi]

grandefuturo,

WOW, what a good find (the ratings), where did you find this? Off the vendors part number, a sticker on the PSU or was it silk screened on the PSU's PCB? The reason I ask is I have reparied about 8 different units and have never seen this.

If this (370mA 9.5 rail) is true, then it supports a theory that Pete (g0pkh) and I have been working on togther on how to improve the PSU for the units that are not built with the update from Roku.

Regards, Matt

[grandefuturo]

Burkhardi,

the PSU's hexagonal board has a rectangular board glued under it.
This lower board is labeled with the data I mentioned earlier.
Possibly, only radios sold in Germany have this label because there's one statement in german "Nur zur Benutzung in geschlossenen Räumen" above the "Made in China". On the other hand, I ordered the radio from London, UK....

Do you know what Roku changed in the updated radios? A complete different PSU? My favourite choice today would be to go with a single voltage supply, possibly a shielded one with metal case, which fits into the space beneath the bass speaker. The 9.5V can be up to 40V according to the CM2596 spec, I need to check the supply tolerance of the audio amp too. If the 19V accepts a high tolerance too, the whole thing can be operated with a single 12, 15,19, or 24V supply.

If you try to optimize the PSU: The display unit takes permanently ~0.5A, even in standby. I've seen in the CM2596 spec, that the switcher has an enable input. Would be great to use this as a real power down mode, with some permanent voltage for the power switch....

[grandefuturo]

I think I got it now.

19V Audio Amp: There are three audio amps, two inside the small stereo speaker cabinets within the display unit (they really are active bass reflex sytems!), and one amp within the woofer/PSU unit. The audio amp's power rail is directly connected to the 19V of the PSU and distributed to the two stereos and the woofer. The audio amps operate without any distortion down to about 5.5V, below they just stop working.

9.5V Microcontroller Voltage:
As mentioned earlier, there's a CM2596 switching regulator down to 5V.
The CM tolerates 7-40V input. However, the cap parallel to the CM input is a 25V only.

Fazit:
There's no reason to have two separate 19V and 9.5V.
The 19V can be down to 5.5V
The 9.5V can be up to 25V

I can confirm this because I operated both voltages with a single 12V,
and the radio was working perfectly, no loss in audio or WLAN receiver performance. The total operating current (audio and microcontroller climbed together to 12V) is about 550mA. At 6V about 1000mA

So I will buy a 12V switching PSU which fits into the bass speaker cabinet. Something not exceeding 110 x 60 x 45mm will fit into the original plastic/alu/copper shielding bricolage, I found some (12V, 30W) for about 25 EUR. Notebook PSUs are too long.

[Burkhardi]

You have been busy…

I think they are Class-D amps with high and low pass filers for speakers and woofer respectively

Some other ideas…
#1 Have a barrel connector on the back of the unit that a 12v wall wart can plug into and use that to power it; I would leave the OEM supply in the woofer enclosure, but not hooked up so that the airspace is not any larger since it’s all a tuned enclosure (with a port). Also make sure the enclosure is sealed up very well.

#2 Use the same barrel connector, but with a 19v wall wart and then have a DC-DC converter inside the unit to go from 19v-9v (probably over kill).

I would guess (since I have not done any testing) that they best/easiest method would be #1; the single 12v external wall wart.

If this does catch on and people like it, then perhaps we could make a parts list to have a harness to go from the 5-pin Molex (in the enclosure)to a barrel connector and/or supply a kit. I saw some pretty scary conversions on the HD1000 with the Morex PSU so this would help curb that.
Then a user could just order the parts from Digi-Key or what have you and be set.

One question that I wonder about is why there was two rails in the first place. I can’t help but think there was a reason that Roku did a 19v and 9.5v rails verse one 12v. Perhaps since it was being designed from the ground up is why (to be more efficient)? I have a nice little 12volt wall wart for an external hard drive in my hand right now, that would fit the bill; it’s not too big, not too heavy and is rated at 12v 2amp. From looking at where you could tap into the enclosure with the barrel conenctor, I would assuem that a wall-wart with a 90degree end would work best.

I may change my tune in 6months or so, but I am going to just fix the OEM supply for the time being.

One last question, how was the AM and/or FM radio portion? Any noise from the 12v supply?

Anyway, great work and this may be a great option once you get some hours behind it to prove there are no strange issues and/or it will hold up (to which I see no reason it will not). Regards, Matt

EDIT: Just remembered there is no molex inside for the DC portion. Let me ponder a tad...

I think I got it now.

19V Audio Amp: There are three audio amps, two inside the small stereo speaker cabinets within the display unit (they really are active bass reflex sytems!), and one amp within the woofer/PSU unit. The audio amp's power rail is directly connected to the 19V of the PSU and distributed to the two stereos and the woofer. The audio amps operate without any distortion down to about 5.5V, below they just stop working.

9.5V Microcontroller Voltage:
As mentioned earlier, there's a CM2596 switching regulator down to 5V.
The CM tolerates 7-40V input. However, the cap parallel to the CM input is a 25V only.

Fazit:
There's no reason to have two separate 19V and 9.5V.
The 19V can be down to 5.5V
The 9.5V can be up to 25V

I can confirm this because I operated both voltages with a single 12V,
and the radio was working perfectly, no loss in audio or WLAN receiver performance. The total operating current (audio and microcontroller climbed together to 12V) is about 550mA. At 6V about 1000mA

So I will buy a 12V switching PSU which fits into the bass speaker cabinet. Something not exceeding 110 x 60 x 45mm will fit into the original plastic/alu/copper shielding bricolage, I found some (12V, 30W) for about 25 EUR. Notebook PSUs are too long.

[grandefuturo]

Burkhardi,

I haven't checked the FM/AM portion, I never used it. I also don't know whether it's connected to 19V or 9.5V.

Your concern regarding the airspace of the woofer: I found a 12V 4A PSU which fits into the original alu/copper shielding box. I also apply the original ferrite beads to the input and output cables of the new PSU, so it should have sufficient filtering also for the FM/AM.

My prefered option is still with the original 115/230V plug on the back, and an integrated new PSU. The new PSU is overload and temperature protected, so I also improved the fire hazard situation.

I will post an update here when I got the new PSU, will take another 2 weeks until I have time again....

[Burkhardi]

Burkhardi,

I haven't checked the FM/AM portion, I never used it. I also don't know whether it's connected to 19V or 9.5V.

Your concern regarding the airspace of the woofer: I found a 12V 4A PSU which fits into the original alu/copper shielding box. I also apply the original ferrite beads to the input and output cables of the new PSU, so it should have sufficient filtering also for the FM/AM.

My prefered option is still with the original 115/230V plug on the back, and an integrated new PSU. The new PSU is overload and temperature protected, so I also improved the fire hazard situation.

I will post an update here when I got the new PSU, will take another 2 weeks until I have time again....

Hi grandefuturo,
RE: AM/FM - I assume it's the 9.5v and that the 19v is just for the amps like you thought.

I perfer the 115/230v plug too and I really like how you where able to replace the PSU with a drop in (per se). Not to mention, you were able to use the OEM sheilding/cage too, which should help with terrestrial radio noise. I was thinking of the 12vDC wall-wart for the weekend warrior that wanted a DIY fix that was not as hard/difficult. Not everyone has the skill set, time or patients. All they would need is to open the enclosure (which is a major pain plus the work area is tight (to do right and w/o damaging the housing)), cut the 5 wires and hook them to the proper parts of a barrel connector that they put thru the backside of the enclosure then seal it back up. I would cut the AC wires from the inside so that if someone plugs it in by mistake to 115/230 that it would not do anything and I would also put the barrel conenctor in the nook where you find the serial number sticker (hence the wall-wart with the 90 degree end). Putting it in the nook will keep it from poking out the rear, make a cleaner install and keep it from being unpluged as easy.

Smart on keeping the ferrite chokes too. I like your style! That is how Peter and I do these, by the book and picky (read over engineered) always better safe then sorry IMHO ))

As for me, I like either fixing/updating the OEM supply or using the exact solution that you did (put another supply in place of the OEM one).

I have a spare enclosure, I may build a wall-wart version just for testing and as an example (like I did with the Morex PSU in the HD1000 post that I did). Plus this will let me check if the audio wattage (for speaker power) is any lower with the 12v feeding the amps and not the 19v. I assume it has a window that will work, but just want to check. I know most people are not going to rock out with a alarm radio, but I would still like to know for my own knowlege if running on 12v will lower the total wattage out of the amps verses 19v.

I know what you mean about having to wait 2 weeks, I have 1yr old twins and finding time is impossible sometimes...LOL

Keep us posted, Matt

[grandefuturo]

OK, the radio is on again. Single 12V PSU, no fault since Friday.
WLAN OK, FM/AM OK, no hum.
AC real power 7.1W / 6.8W (Standby)

[Burkhardi]

OK, the radio is on again. Single 12V PSU, no fault since Friday.
WLAN OK, FM/AM OK, no hum.
AC real power 7.1W / 6.8W (Standby)

grandefuturo,
I have had mine on an external wall-wart (12vDC@3amp) since Friday and it seems to have good WLAN, FM and no hum/noise that I can hear. I have not tried the AM yet since I didn’t have the antennae handy; but I will check it out. Worse case I will finish this up over the Christmas holiday. I am running this test on my personal unit and once I am done, will be repairing it's orginal power supply as well. BTW, the above test was done with the original power supply still in the enclsoure but not hooked up, so the airspace for the woofer would be correct.

My next step is to put it on the bench and see how the output (audio amp wise) compares to a unit in the normal configuration, as well as checking for clipping and/or distortion at high outputs. I also want to use the bench power supplies and see the current draw under a real load to double check to what the required amperage should be. I want to check them in the OEM setup (9.5vDC and 19vDC) and then combine them with 12vDC feeding them.

I was able to repair two more units over the Thanksgiving holiday (repaired the original power supplies); and I replaced the caps with good quality Japanese ones and they ran great all weekend too. I will be sending those back to the owners today.

Thanks for the feedback, Matt

[skydreamer]

My R1000 power supply also failed just before Christmas and a simple replacement of the blown up 2,200 uF caps did not help hence I ripped out the power supply and connected the +9 and +19V rails to the 12V switching mode PSU from Maplin. So far so good although I have noticed some early base limitation at the higher volume levels.
The 4 Ohm base speaker is able of theoretical output of 36 Watts at 12 Volts or 90 Watts at 19 Volts although guessing by the size of the speaker the sustained acoustic power is likely to be closer to 10 Watts so 12V power supply is plenty.
On another note removing the PSU is not going to change the acoustic qualities of the base cavity dramatically since the PSU enclosure is made of a flexible material and therefore does not affect much the resulting sound. The perfectionists can replace the PSU with some extra sound dampening material.
I did not connect the shield wire for fear of creating a ground loop and the radio works perfectly fine.