Phonic Firefy 808 Issues

[Stef123]

Hi,

sorry for not introducing myself first - i have a quite delicate Problem with my Phonic Firefly 808 Universal (Firewire/USB) from 2012, which i bought in non working condition and getting Schematics / Service Manuals for this thing, as well as for other devices based on the BridgeCo DM-Series (DM-Series was also utilized by Terratec, M-Audio etc, any Audio Interface which incoorporates Firewire)), looks very impossible -idk why, because this Platform and technology around it is quite outdated.

Well, it was said by the seller that it cannot be recogized by USB any longer, and a close inspection revealed that the D+Trace of the USB Port was "flying" around; no idea how that could happen. Means that i had to make a new connection, which i eventually did. But it looks like as the Firefly "resets" for each two seconds, because the Power light dims down for a short amount of time - before it does that again in a repeated loop.

Oscilloscope shows that the Voltage on the CPU PCB drops for a short period, but not on the serving Main PCB. So hooked up a logic analyzer to something "what seems to be a 4-Pin debug port" - no idea which Format it has, i suppose serial, and it outputs something, but repeatedly for each two seconds.

USB D+ and D- is low all the Time, even with a computer connected, which in conclusion cannot be right. They go straight into the Main DM1500 ASIC, which seems to boot a stripped down Linux Kernel from Flash. Of course there is also no Datasheet of that ASIC anywhere available, because the Company, which bought BridgeCo around a decade ago, closed Business at the very beginning of 2023.

So the question is: Whats happening here. As for my own experience of programming Microcontrollers for +30 Years, it -might- be possible that the integrated USB port of the ASIC changes State from unconfigured to something that does not "quite" like 3.3 / 5 Volts being pulled directly to GND in case of a defect, but i assume that should happen right after power up. And the ASIC itself won´t get hot at all, but seems to run (otherwise i wouldnt get "debug data" out of it, in whatever format it is). A diode Test revels at least no short curcuit between GND and the USB D-Pins.

Any help - and probaly someone which can share additional information about that specific Platform, at best a Service Manual for the Firefly, is very welcome!

Best regards!

 

[Stef123]

Weii, in my Opinion the TIP41C (65 Watts) doesn´t look too good at all. A closer Inspection revealed that one ceramic capacitor was more or less burnt away, i guess by the heat of the power resistor. Additional Components on the "low voltage"-Side are some SR840/880 Rectifiers and a TL431 100mA Voltage Regulator.

 

[rjenkinsgb]

The solder joints on the large grey resistor just in front of the capacitor look rather dubious?

 

[Stef123]

Yup, but has good Contact on the Bottom Side.

Well, the TIP41 and both other BJTs are, as it seems, not responsible for this issue. The TIP41 provides +48 Volts (for Microphone Phantom Power, i suppose), the other two on the coolig element roughly -15 and +15 Volts for Opamps. So the +5V must have its Source near the Transformer. The virtually only thing responsible could be the TL431, but he would be way too weak for the power needs of the main PCB and some other 5V Parts. I cannot imagine that a 240 MHz ASIC, RAM, ROM, Firewire Chip and so are happy with 100 mA @5V.

 

[Stef123]

If they -have- to be happy with it, then the problem gets worse, because it would mean that the main PCB draws more power than available- hopefully not because of the defect USB port. The only thing i can try then is to cut USB Power on the ASIC to deactivate it.

 

[rjenkinsgb]

I'd have thought the 5V rail was likely a main output from the switch mode supply, probably the one the regulating feedback is taken from?

 

[Stef123]

I suppose you are right, I´m not into switching power supplys. The TL431 only plays a role in Voltage Regulation, but without,as it seems, without feeding a Power Transistor, as shown in various examples. Since Voltage is dropping on load, i guess he cannot keep up.

Thing is, as measured, the Voltage drops from roundabout 5 Volts roughly .2 Volts to 4.8 Volts, while taken Amps rise from 100mA (the point where the ASIC starts to verify the loaded image) to 200 mA and the Power LEDs get slightly darker - which is the point where he wants to finally boot - and then the usual reset of the ASIC occurs.

Thats something i also experienced with the CPU Board connected to an external PSU. If it gets power from a typical LM2596 DC Buck Converter, which precisely doesn´t like fast load changes, the same error occurs, but not, if the CPU Board is fed directly from a PC PSU, which can keep up better with fast changes in Power requirement.

 

[Stef123]

Short story: Idk which component might be responsible for the inability to keep up with the rising load and how much it should provide by standard. I can only suppose that the ASIC sees an undervolt condition and therefore resets. Edit: Wrong measurements. Drops from 5V to 4.3 Volts. Thats clearly undervolting. "Control Voltage" of the TL431 is 2.2 Volts at 4.3 Volts output and starts rising under non-load conditions.

 

[Stef123]

Success! Changed a 600uf/10V-Cap, which only had 515uf left, with a 1000uf/10V and a 47uf-Cap near the TL471. Both used, but working. Result: Better change all Caps.

Well, that was '"some kind of birth". Thank you for your support !

 

[Stef123]

I/O is also working through ALSA (preferred over JACK), altrough it can only use max 96 KHz of 192 KHz possible . What will be added for sure is fan cooling. That thing has virtually no heat exchange to the outside world, only through the Metal frame - inacceptable for long lasting.

 

[rjenkinsgb]

The main incoming power cap (350 or 400V rated?), just to the right of the transformer in the photo, is also worth checking?
There is some discolouration near it which could mean its oozing electrolyte?

 

[Stef123]

I´ll take a look; idk if this is remaining Flux (however it got there) or oozing. Eventually i will change every electrolytic at least on the PSU side, but i have to order them first with the next order from my my electronics store. I have most of them as 16V Types, but it uses mainly 25V Types. Could be usable in the 5V-Rail, but the bigger, the better.

 

[Nigel Goodwin]

I´ll take a look; idk if this is remaining Flux (however it got there) or oozing. Eventually i will change every electrolytic at least on the PSU side, but i have to order them first with the next order from my my electronics store. I have most of them as 16V Types, but it uses mainly 25V Types. Could be usable in the 5V-Rail, but the bigger, the better.

I would suggest it's most likely just adhesive?, holding the large reservoir capacitor in place.

Also, I wouldn't get too bothered about changing too many capacitors - just the ones in the secondary of the PSU.

The ones you changed weren't 'low capacitance' they were actually high-ESR, and identifying them is what ESR meters are for.

It's been the most common fault in electronics for a good many years now, and an ESR meter is a service engineers best friend If you've not got an ESR meter, but do have a scope, (and the PSU is running) then checking for ripple on the secondary rectifier diodes outputs (easily identified) will show huge ripple on rails with high-ESR capacitors.

Basically cheap capacitors commonly fail in switch-mode power supplies, where they are used in high frequency applications, so mainly all the secondary reservoir capacitors, such as your 600uF and 1000uF.

The large primary reservoir capacitor only rarely fails, as it's only operating at 100Hz.

I've changed thousands (probably tens of thousands?) of high-ESR electrolytics - if you're lucky that's all that's needed, but sometimes it causes more damage as well.

 

[Stef123]

Yes, an ESR Meter will be indeed useful. I only have a small and rather old 10 MHz Analog Scope, a 20 Year old but still useful Fluke Meter and some other stuff like a 16 Ch Logic Analyzer. Doing electronics stuff for about 30 Years, but mainly "less complex" designs (up to double Layer 160x100) but more complex ASM-programming with Microcontrollers.

 

[Nigel Goodwin]

Yes, an ESR Meter will be indeed useful. I only have a small and rather old 10 MHz Analog Scope, a 20 Year old but still useful Fluke Meter and some other stuff like a 16 Ch Logic Analyzer. Doing electronics stuff for about 30 Years, but mainly "less complex" designs (up to double Layer 160x100) but more complex ASM-programming with Microcontrollers.

A 10MHz analogue scope is perfectly fine for checking for high-ESR capacitors, there's no advantage in using a higher spec. or a digital scope. You'd just looking for something like 5V p-p ripple at 50KHz-100KHz on the output of the rectifier, where it should be 100mV or so.

So scope requirements are very low - it's not like you even need to accurately measure it - just BIG or SMALL

 

[Stef123]

Yes, i´ve seen something on the 5V Rail when i measured it in order to look for some weird behaviour which couldn´t be captured with the Fluke , but it was more Noise in the mV-Area rather than ripple. Real Peak-to-Peak Ripple in order of Volts was not visible at all.

 

[Nigel Goodwin]

Yes, i´ve seen something on the 5V Rail when i measured it in order to look for some weird behaviour which couldn´t be captured with the Fluke , but it was more Noise in the mV-Area rather than ripple. Real Peak-to-Peak Ripple in order of Volts was not visible at all.

The 5V rail may be too far away - there's often a rectifier, with a reservoir capacitor, then a choke, and a smoothing capacitor - you need to measure on the output of the rectifier (top of the reservoir capacitor) - that's where the largeripple usually is. If the reservoir is high-ESR, the choke and smoothing capacitor will remove most of the ripple from the actual rail itself.

 

[Stef123]

Thanks for that very useful information ! Yes, the "rather large" smoothing Cap with faulty readings was indeed that one which has been replaced. Means that the reservoir cap could be still faulty - but as soon as all of the 2nd PSU side caps get replaced, this should be no issue no more.

 

[Nigel Goodwin]

There were 'service kits' available for numerous satellite TV boxes, which basically included all the electrolytics required to replace the ones that commonly failed - we used to keep multiples (fives or tens) of the various different types. I also built my own 'kits' for repairing various TV's - often the PSU was destroyed by failure of the electrolytics, and I assembled kits for my own use that included all the usual parts that failed (could be 20+ parts in some cases?.

When you know what fails, and what's likely to fail, it's much more cost effective to just do a blanket change of all likely parts, rather than spend expensive engineer time diagnosing exactly which parts have failed in any particular case.

One 'unusual' point - the large mains reservoir capacitor - those do occasionally go faulty (although it's not that common), and on some sets this causes the PSU to go 'bang' Yet different makes of sets, using the same PSU IC, and pretty well identical PSU circuitry, survive, and even continue to work albeit with a hum bar on the screen. I could never quite understand why?.

 

[Stef123]

On my other specialized field, repairing analog Synthesizers and vintage 8/16 Bit Computers, there are also often some cap Kits available for PSU and analog circuitry. But in terms of the PSU i prefer to exchange the entire PSU with a modern type (expecially if they use "heaters" like 78-Series, as it was normal back in those days), if something is available that fits. Commodore Brick-PSUs for example are often filled with epoxy, so there is anyway little chance of repair, althrough it can be removed with things like Dichlormethane, but that is a mess and a repair doesn´t make it lasting longer.

TV-Repair sounds quite interesting - my repair abilites in this field are restricted to change the caps of the LCD-backlight, if it fails (usually they are undersized in terms of Volts). For CRTs i better enjoy Shango066s Videos when he tries to get various old TV sets working again, because i have some serious respect from them.

As a kid, i also had often my hands on my Parents B/W-TV when they were not at home, wondering whats happening if i pull the red cable off the tube while it was running; adding my Recorder to the deflection coil in order trying to record something and so on Christmas / Birthday electronics presents usually didn´t last longer than one day, because the innards were more interesting....

 

[sbrendtro]

I/O is also working through ALSA (preferred over JACK), altrough it can only use max 96 KHz of 192 KHz possible . What will be added for sure is fan cooling. That thing has virtually no heat exchange to the outside world, only through the Metal frame - inacceptable for long lasting.

First of all, THANK YOU! I've been looking for a solution to this behavior for 5 years now when my FireFly 808U started cycling itself. Interestingly, after a year or two of sitting in the closet (and a new version of Ubuntu), it seems to have stopped cycling for the moment (self-healing capacitors?!?) I'm going to replace the caps now for good measure.

I do have another issue that you seem to have gotten past. (Sorry if it's slightly off-topic from hardware/electronics... this is software.) My FireFly is stuck at 44.1KHz, and nothing I can do will change it. You mentioned Jack, so I assume you're on Linux as I am. If I set Jack to use a higher sample rate, it stays locked on 44.1K (on the unit's LED), and causes a distorted mess of audio (from the mismatch, no doubt). Did you do anything special to get it to 96 KHz? I can't figure out how to tell the FireFly to switch rates. Windows had the ASIO utility that allowed you to switch, but I can't figure it out.

Thanks again for the deep dive into the FireFly!