It looked like a regular smd XO, with 10 MHz going straight into the fpga, but I could be mistaken. I only opened it up to have a quick glance at the internals before daring to plug it in. This turned out to be a good idea, because the cover screw had inconveniently drilled itself straight through the SMPS ribbon cable, pretty much shorting together the +-13.5v rails!
Other things I noticed were the NEC relays, with date codes ranging from 1997-2004. Three of the four had what looked like some kind of conformal coating on them, so I suspect these werent even NOS relays 
Regardless, still a bargain, but quite hilarious. I'll have a closer look at the main board next time I have it open... when I replace the power supply.
Just when you thought they couldn't get any cheaper! Their buyer (buying department) must be expert(s) at sniffing out bargains in the component supply chain. Perhaps they'd located a dirt cheap source of 10MHz XO and clock multiplier chips (not necessarily those NB3N502 ICs I mentioned) or perhaps an FPGA version that can be clocked with 10MHz rather than the more typical 50MHz - who knows how the inscrutable Chinese mind works?.
Incidentally, are those +/-13.5v rails marked as such, or was it just what voltages you were seeing whilst the unit was delivering 20Vpp into 50 ohm dummy loads? Also, did you take the opportunity to repair the vandalism Feeltech had perpetrated in stealing one of the two ground return wires out of the ribbon cable in order to provide a
hard connection to the PE terminal on that BFO C14 mains connector?
This latter act of vandalism really ought to be remedied, preferably by tripling up the one remaining ground return wire and patching out the missing wire similarly and, for good measure, remove the resulting ground loop issue by inserting a 3K3R or 10KR in series with the PE connection. The class II smpsu board used doesn't actually require a hard protective earth connection and even a 10KR "static drain" resistor is sufficient to knock the 90vac half mains live voltage ESD hazard of a 240v mains supply down to just half a volt ac or less.
At least one member who'd bought an FY6800 was saved the trouble and embarrassment of returning it with a "DC Offset fault" after heeding advice to break the earth loop (in his case, he'd chosen to disconnect the 'scope ground) and retest. The FY6600 hadn't suffered this ground loop issue but the half mains live 'touch voltage' presented an ESD hazard if you forgot to connect the BNC ground
before connecting to the test point of your DUT (and disconnected it
after disconnecting from the test point).
It turned out that the solution was simply to upgrade the 2 pole C8 connector to a 3 pole C6 or C14 connector to wire the PE to the main board ground via a 10KR 'drain resistor' to suppress this half mains live 'touch voltage' and neatly sidestep this ground loop issue.
It's quite clear that Feeltech's decision to upgrade from the C8 to a C14 mains connector on the FY6800 model must have been a response to all the moaning about the ESD hazard in the FY6600 EEVBlog thread. Unfortunately, they chose to do the absolute minimum possible to implement a solution to address this complaint about the presence of this ESD hazard by totally ignoring all and every "unintended consequence" of hard earthing the BNC ground points and reduction of ground return wires in the smpsu board to main board ribbon cable connector from a barely adequate count of two to a woefully inadequate count of one.
That ribbon connector could have done with an upgrade from a 6 pole single row to a 12 pole dual row connector to add another 6 ground wires to the circuit to reduce PSU noise and ripple getting onto the main board supply rails. Reducing the ground return wire count would have been the last thing any sane manufacturer would have even considered, let alone implement.
Feeltech (FeelElec) aren't the only Chinese manufacturer of bargain priced test, measurement and electronic tools who show such blatant disregard for any modicum of quality in their products. KSGER who make all those otherwise excellent T12/T15 tip based soldering stations are just as bad (and even worse if you rate the electrocution risk higher than the ESD risk).
I bought one of these cheap alternatives to Hakko's rather flawed FX-951 soldering station two months ago and, as with the FY6600, it rather benefited from some basic fettling and general remedial work. It is now as safe as any normal electrical appliance made by The West (excepting for some models of Weller soldering stations
) can be expected to be and no longer at risk of its controller succumbing to soldering iron handle wiring faults, typical of "Chinese Quality Control".
It's considered SOP to take such mains powered Chinese Marvels of Technology apart as you did, in order to check for and (almost inevitably) make safe for "Human Consumption" and 'fitter for purpose'.
I was only expressing an idle curiosity in the matter seeing as how Feeltech appear to have been taking note of the various moans and groans and modifications that had been discussed in the earlier FY6600 thread (a "Standing on the shoulders of giants" effect) where the use of the 3N502 had been suggested by one of the members to Arthur Dent who had been proposing to fit a 10MHz OCXO to his FY6600 at the time (and has long since done so) and from which I eventually[1] took inspiration to do likewise after having already replaced the shitty little 50MHz smd XO chip with a 50MHz 0.1ppm rated TCXO driver board I'd mounted well clear of the 50 deg C environment of the original on-board XO chip.
Having employed the 3N502 to allow me the use of a CQE branded 12v 10MHz OCXO in my own FY6600 project and using another one in my current DIY GPSDO project to make use of a 5v 13MHz CQE branded OCXO (74HC86 clock doubler to 26MHz into a 74193 (original old skool TTL!) divide by 13 to drive the 3N502 with the lowest allowable 2MHz input frequency clock) to generate an ultra low jitter 10Mz square wave, that 3N502 was rather on my mind when I saw your post.
As I've already admitted, it was just idle curiosity on my part so there's no hurry in answering that question. That's not to say I wouldn't be interested in what you do eventually discover as I'm sure others would likewise be interested in what you may care to divulge about Feeltech's (FeelElec) latest and greatest toy signal generator.
[1] I rather naively thought at the time, that this OCXO mod was just a tad OTT but eventually realised the value of such accuracy and stability once I discovered that I was never likely to improve on the +/-30ppb I was getting out of the 50MHz TCXO board. As good an improvement as it was over the original smd XO chip's best +/-20ppm effort at (in)stability, I had developed a new found desire for sub ppb accuracy which nothing less than an OCXO could satisfy.
JBG
Home
Help
Search
Login
Register
