Ok, time for some nostalgia. My first "big" interface (as in, not just a little 2in/2out jobbie) was a Saffire PRO 40, that i bought retail, brand new, back in 2011 - spring, if memory serves. I mostly used it as an interface at home, tracked a couple rehearsals with the band i was in at the time, then later it got "demoted" to additional-preamp to a Saffire PRO 26io that i briefly owned, before it got sold off when i got my MOTU 896mk3 Hybrid.
This one i grabbed as faulty off eBay, yet again. Another one of those "too cheap to pass up" sort of deals. The listing contained a confession that the seller had (unsuccessfully) attempted to replace the Firewire chip, and he also provided a photo of the affected area. I won't lie, it kinda hurt a tiny bit to see a couple of torn off pads, and a few others lifted and mangled, but at the same time, it also didn't look to be entirely beyond hope. I figured with some determination and a bunch of patience (for "cosmetic / reconstructive surgery"), it could be returned to full functioning.
Plentiful enough ins & outs on this lil' feller - continuing from the two mic/line/instrument inputs on the front, the rear of the unit hosts the other six mic/line inputs, as well as the ten(!) line outputs, the optical in & out, the two Firewire ports (which have seen better days), MIDI in/out as well as S/PDIF in/out.
Since the seller kindly(?) shipped the unit ready-disassembled, but with all(?) the screws in a plastic baggie, popping the lid off is virtually instant. What is revealed will not be much of a surprise to those of you who have looked into the teardown of the later Scarlett 18i20. Since in that unit, they hadn't even bothered to change the silkscreen on the front panel PCB, the similarities in design, layout, component choices etc. should shock noone.
I wouldn't blame you if you could barely tell this apart from its... "nephew" (or "niece"?) unit. I wouldn't be at all surprised if they just copied the design documents and just altered whatever they needed to, in the digital / processor areas. Work smarter, not harder, eh?
Might as well start with the power supply. Apart from the date code (week 29 of 2013, as opposed to week 10 of the same year), and the brands of capacitors, the power supply is exactly the same. The switching MOSFET is a CR Micro(?) CS8N60F, while the diodes on the secondary are a pair of ER304 for the opamp supplies, an ER10-something for the phantom-power rail, and a... double diode i can't read the markings on, for the digital supply rail. Probably some not-too-remarkable 2x10A fast-recovery diode (best case, a Schottky), in any case. The primary cap is branded "Tianchen", while the secondaries are... [shudder] CapXon...
The voltage regulation section on the mainboard is comprised of an LM317 / LM337 pair for the +/-7V for the various opamps, and a second LM317 providing a clean 48V for the phantom-power.
The major difference between the Scarlett and the Saffire is the main processor used. While in the newer unit there's a pair of Xmos chips, in this case the heavy lifting is handled by a TC Applied Technologies TCD2220 "Dice Jr". Exactly the same chip as in the M-Audio ProFire 2626, in fact. That is assisted by an ISSI IS42S16800E-7L 128Mbit / 16Mbyte SDRAM chip, and the firmware is contained in a Spansion S29AL016J70 16Mbit / 2Mbyte Flash memory. And the silver spot is where a TI TSB41AB2 was supposed to be, but as hinted in the intro, that had kicked the bucket. Further details to follow below.
Yyyyyyeah, this board has definitely seen better days... Three pads missing - one's number 16 on the bottom edge, but that's "not connected" and thus irrelevant; the bottom-most one along the left edge going to ground (but its neighbour is also a ground pin, and a little solder between them should do the trick); and the one on the top edge used to also be connected to ground. Fortunately there's a soldermask-less via in the groundplane right next to it, so it shouldn't be too-too much of a headache to salvage that. Other than that, at least a quarter of the remaining pads are lifted - ouch. That's gonna take some surgery-ing...
In the mean time, moving on to the inventorying of the silicon - once again, the audio conversion is performed by a quartet of Cirrus Logic CS4272's, with an additional CS4392 DAC for the "main outs" (line outputs 1/2).
The preamp section for the two front inputs (mic / line / instrument) is handled by a pair of NJM2122's as the preamps themselves, a couple of HEF4052 and HEF4053 for the switching / routing of signals, as well as a pair of TI TL072's for the instrument inputs. The resulting output signals are buffered by an NJM4565.
The other six analog inputs have further NJM2122's, HEF4053's and NJM4565's associated with them.
A boatload of NJM4565's buffer the analog ins & outs of the converters.
And a few more NJM4565's drive the line outputs. Also, the main stereo output also goes through a pair of Matsushita / Panasonic TQ2 relays, for signal muting (power-up delay).
The S/PDIF in & out are handled by a National Semiconductor DS89C21 line driver / receiver.
The two headphone outputs are driven by an NJM4556AL each.
Now then, i mentioned a bit earlier this would require a bit of surgery, what with the mangled state of the footprint for the Firewire interface chip. A bit of added flux and fresh solder, followed by careful dabbing with solder-wick resulted in a promising step forward, if i may say so myself. The missing ground pad was "reconstructed" with a carefully placed / installed strand of copper wire, with one end bent and carefully inserted into the nearby groudplane via.
I also did my best to carefully straighten out and realign most of the loose pads. Fortunately, no further ones were lost or broken off. A few minutes of careful soldering later, and...
If anything, that went even less dramatically than i would've expected. With that done, and after making sure the 3.3V rail wasn't shorted to ground, i flicked the power on and... Nothing blew up. So far, so good. After installing the appropriate drivers on an old test-laptop, Firewire got plugged in and... it works!
Well, in all fairness, not flawlessly, not straight away. Something must've gone astray in Windows' audio clock settings somehow - the unit seemed to cycle between various sampling frequencies (if the numbers displayed in MixControl are to be believed). But after digging into the old Control Panel, opening the Sounds applet and selecting 24bit 44.1kHz for the Saffire audio device, it settled down, the "FW Active" LED stayed solidly on, and all was well. Analog ins & outs test fine, ADAT in/out looks good too, so... Score another one!...
I have saffirepro40, first unit turn on,psu voltage drop down after 10minute, so this unit restart after 10 minute, then always restart about 2minute.
ReplyDeleteCan you help me??
First step would be to replace the electrolytic capacitors in the power supply, with some proper ones (Rubycon, Panasonic, Elna, Nichicon or United Chemi-Com) - low-ESR, of course. The 400V one on the primary doesn't need replacing, though.
DeleteOnce that's done, if it still behaves the same way, then it's time to dig further into what might be the root cause.
In the mean time though, i could point you towards the thread i started on the badcaps.net forums many years ago, about this model in particular. There may be others who can offer helpful advice there.
https://www.badcaps.net/forum/showthread.php?t=22485
Nice Project. I fixed my 2007 Vizio TV a few years back thanks to a youtube video and a $2.00 12V power transistor, so i googled sapphire repair and found your blog and here I go again looking for a cheap way to fix my old sapphire 40.
ReplyDeleteI bought My Sapphire pro 40, the same year as yourself @ guitar center for $499. I'm using adat in/out to connect it with my new rme raydat pcie card and mixer software. It worked for months, great low delay @1.2 ms. But Recently I noticed Its missing all channel 4 volume level indication on my rme mixer's gui software. And I also can't hear chan 4 in my rme raydat mixer [via that computer's audio chip] or the adat return path back to my sapphire 40 @ch[9, 10] headphone out. Channel 4 vol level indication is present on the sapphire 40 front panel leds and I can hear chan 4 in my sapphire 40 headphone monitors @ch[7.8]. The Channel 4 vol level indication is present on the sapphire mixer gui when I connect via fire wire on a separate computer which i use to reconfigure the sapphire mixer hardware with preset hardware files for mixer/adat connections from focusrite online tutorials. the config is known preproven good.
So wo opening the lid, i think i can deduce there is a problem somewhere in the sapphire 40 chan4 dac to adat digital injection path. I'm not at all familiar with the signal path from dac to fire-wire and adat output?
As a retired EE I would bill my time at 50.00/hr so anything over 10 hours I consider a loss unless it's fun or not too frustrating. Otherwise It serves my old brain better to buy a used working unit rather than get too frustrated by a difficult repair. I hate when i get started and then cant fix something.
So, if you've read this far and choose to humor me, my question would be if you have enough familiarity following the dac paths and if so could you narrow down my potential repair items by these symptoms or is taking off the lid not worth my time. I figure you have more insight on this item. since you've had the lid off and already had a decent look at the signal paths?
OOPS
Deleteon the focusrite mixer router
i had
ADAT4in ---> ADAT4out
should be
ANLG4in ---> ADAT4out
i think i downloaded a bad tutorial set up
AS ALWAYS CHALK IT UP TO PILOT ERROR - BAD EYES
The only thing between the analog-to-digital converters is the main controller / processor chip, the TCD2220. One single ADAT output channel missing is something i hadn't heard of before. But if there's level indication on the input meters both on the front panel and in MixControl, the input side of things should indeed be fine. And the ADAT connections go straight into / come straight out of the TCD2220, so whatever fault (if any), will be internal to that.
DeleteIf you're willing to take the chance, you can find those on AliExpress for anything between $10-40. I've recently bought three; used one to revive a ProFire 2626, and one for another Saffire 40 (that i bought as faulty, but was missing the TCD2220, the RAM, the flash memory and the Firewire chip; this one still has a power supply issue as well, though).
Hello, hopefully you could help me ! The psu of my pro40 just died few days ago. I cannot read the resistors labels (backside of the psu card) as they burnt. Can you help ?
ReplyDeleteI'm going to assume you also posted on my thread on the badcaps.net forums? If yes, i've replied there.
Deletethank you
ReplyDeleteHi There --
ReplyDeleteI've got an ol' pro40, and I miss having it! The input lights work, showing its receiving a signal -- and the computer recognizes it -- but no audio comes through.
Any ideas about a DIY fix? Many Thanks!!
It seems I've learned I could send it to you rather than chuck it :)
That's quite an interesting symptom.
DeleteAre you 100% sure the MixControl routing has signals / mixes going out to the outputs you're expecting audio out of?
I suppose it could conceivably be the converters, but that's somewhat doubtful. I could be wrong, but i'm reasonably sure the level meters are driven from the digital side of things, and the converters are "combined". As such, if some clock signals were missing, for example, the analog-to-digital conversion would be dead too, not only the digital-to-analog side.
Ideally, you'd want to use an oscilloscope to probe the signal path at least from the DAC outputs, all the way to the output jacks. Although, of course, it would be safe to also "sanity check" the clock signals being fed to the converter chips, but as i said, i would suspect that side of things is fine.
Hello,
ReplyDeleteI have my Pro40 for 10 years now, powered on nearly every day 24/7.
Yesterday all of a sudden it stopped working, and doesnt power on since then. I opened it, something smells burnt but i cannot find where the problem is. I have zero experience and knowledge with fixing things and in Greece i cant find a good place to send to fix it. Any suggestions what it might be the prolem?
I'd start with replacing all the electrolytic capacitors on the secondary side of the power supply.
DeleteBut even before that, have you taken a look on the underside of the power supply? Something might be toasted there, without showing on the top side
Hi, thanks for sharing this. I've been very unlucky lately with these Saffire Pro 40 units. First one got fried by an power spike in my neighborhood... so I really did all I could to fix it, managed to buy a replacement PSU on eBay, paid serious money to get it delivered to Brazil (where I live) but even if the unit got to power on again it did not fix the firewire or something else which also got fried. So I got another working unit, all good, and had it happily working for a week... until the firewire 400 to 800 cable disconnected from my old Mac Pro and I managed to insert the 800 port turned 180 degrees to the computer. I could hear something getting fried and now the firewire ports on the Mac do not work anymore, neither the interface will connect to any other computer. Probably the firewire chip. Again. So now I have these two identical faulty units and as I really need them to work I am going to buy a THIRD ONE even if I hate this idea. Anyway I took it to a Focusrite authorized repair shop and the guy said he doesn't have a special soldering machine to replace this chip. But it looks like you did it manually on your unit. So i'd like to ask: do you believe I could find a place to replace these faulty firewire chips somewhere else in my country, or somebody who could do it manually? And where to get these firewire chips? Thanks
ReplyDeleteThat's very unfortunate indeed. I never knew you even could plug the 800 connectors upside down (at least not as easily as the 400's).
DeleteRegarding the FireWire chip replacing - it's easiest to remove with a hot-air station, but i do prefer to solder in the new one manually indeed.
I've read about one option being cutting the legs off a chip like that (but reeeeeally carefully so as not to cut through any PCB traces underneith). Unfortunately that's not really an option here, since they DID solder the belly-pad of the chip to the ground plane, so it can't really be done without hot-air.
And i prefer the get my TSB41AB1/2's from reputable suppliers (like Mouser, Digikey, Farnell etc), to minimize the risk of counterfeits.
Hi Khron,
ReplyDeleteExcellent work for a nice easy repair. I have a Pro 40 with an irritating issue that took a few days to diagnose. It seems that when using phantom power, after about 3 minutes, playback will start to crackle a little. And then more, and more, until it is a digital mess. I have tried 1-4 and 5-8 and the issue persists. Any ideas of what may be causing this?
Kind regards,
Mitchell
... But it consistently stays sounding ok, if the phantom power stays off? Is the issue presenting (only) on the analog outs, or the digital ones as well? And/or is it audible (also) on recordings from the analog inputs?
DeleteHello, I recently purchased a Saffire Pro 40 but unfortunately it came with one of the headphone potentiometers broken (the shaft). I opened it up and I could see 5421 and 103B on the pot, which, if I'm not wrong is a linear 10K pot, however I'm not sure about the type of potentiometer because it has 7 pins, is that an encoder? And do you perhaps know where to get these from? Thanks
ReplyDeleteThe -A suffix usually stands for linear taper; this is more likely a log-taper one (especially since it's a hardware volume control). It's "just" a stereo pot (with 3 pins per each of the two internal pots), with a spare pin for mechanical purposes.
DeleteBest bet would be one of the bigger component retailers (TME in Poland, Mouser & Digikey in the US, Farnell / element14 on both sides of the pond). Look for a 10k stereo / dual potentiometer, filter by "in stock" only, optionally sort by ascending price, and see which one looks right. Once you've found that, double-check the dimensions with the drawings in the datasheet, as well as the exact part number (there could be several options for shaft diameter, style, length etc).
Hi Khron, thanks for sharing your knowledge and experience, it's really helpful. I have an issue with my saffire pro 40, on channel 5 the potentiometer doesn't do a thing, I've plugged a line level signal and even with the potentiometer at maximum it doesn't increase the gain at all, the same for a dynamic mic with an XLR cable, it does get signal, but pretty low, I have to increase the gain before the preamp or shout in the mic to get the mix control to show any level. From what I remember it's been like that since I bought it, but ignored it thinking it was the source's fault, or the mic not working, etc. Do you have any idea of what could cause that? Thanks
ReplyDeleteIt might be worth popping the cover off and see if the cable between that gain pot on the front panel, and its respective preamp, is still in one piece, and the connector seated properly. Especially if you bought it used.
Deletea bit of a long shot, but does anybody know what kind of inductors are used on this motherboard as well as their value?
ReplyDeleteAre any of yours damaged, or..?
Delete