Re: FeelElec New Arrival FY-6900 Signal Generator

[Labrat101]

Better still operate only while standing on rubber mat .
 ⚠️⚡⚡

[Johnny B Good]

 Feeltech bodged a "safety earth" connection on the later 6800 and 6900 models (replacing the two pole C8 mains socket with a C14 and stole one of the two ground return 3A rated ribbon connector wires to divert to the safety earth tag on the C14 socket) which added other dangers beyond the half live 'touch voltage' of the FY6600 on the BNC grounds which is merely an annoying irritation to human life but a serious ESD risk to any sensitive DUTs due to the stored charge on the Y capacitor that curses all EMC compliant class II smpsus such as the 10W rated one used in all of these AWGs.

 The later 6800 and 6900 models didn't suffer from this 'half live touch voltage' issue since Feeltech had swapped it for an even more serious issue of fire risk as well as providing a path from the polluted earth wiring for noise and dc offsets to interfere with the AWG's output signal, a defect that hadn't existed in the 6600 and earlier models.

 With the 6600, the optimum solution to solving the ESD risk (and accompanying half live touch voltage annoyance) turned out to be simply upgrade to a C14 or C6 three pole socket to gain access to the safety earth by which to link the BNC grounds via a 1 to 10 KR resistor to kill off the touch voltage ESD risk, reducing it to 250mv ac or less without providing a lo-Z path for switching noise and DC offsets from the mains wiring as well as switching noise from its own smpsu getting back into the mains supply.

 The class II smpsu boards in all of these models simply don't require a safety earth connection so replacing the grounding wire link used in the later models with a 1 to 10KR resistor does not compromise electrical safety any further than the regulatory bodies involved in setting such standards already allow with the "Y cap bodge" used in low power (<75W?) smpsu to meet their EMC directives.

 That FY6900 you're contemplating on purchasing, assuming no manufacturing defect , doesn't suffer the 'touch voltage'/ESD risk issue that had plagued the earlier FY6600 model range. However, it will suffer random DC offset and noise injection via its Lo-Z mains earth connection and, worse still introduce an elevated risk of conflagration due to the way Feeltech had bodged their earth wiring. On top of which, there remains the small risk of an attached DUT developing a full live contact fault with its own grounding point.

 That 3A rated ribbon connector conductor used by Feeltech might be able to blow a 2A glass fuse to smithereens but it's woefully inadequate for the task of taking out a 13A plugtop fuse (UK case) that the DUT's own safety earth connection is meant to blow open circuit in a timely fashion.

Admittedly In most cases, such events are more likely to result in nothing more than the release of some 'magic smoke' with just some charred pvc insulation hanging off of a much abused ribbon conductor as evidence of the event but, when your life (and that of others) depends on such fuse based fire protection measures, I wouldn't bank on such a version of Russian (Chinese) Roulette always ending quite so benignly.

[john_mallord]

FY6900 just arrived this morning. I have attached some pics.

Looks like the PSU is new. Very different from the PSU's of the various reviews in youtube. This time, its only using a single output of 5V.
Mains wiring are now with insulated terminal lugs compare to a nasty soldered wires before. But seems like the PE is still using the same thin wire compare to the much common thicker green/yellow wires for earth wiring.

[Janwoj]

FY6900 just arrived this morning.....

I am interested in whether it is the FY6900-100M model or another, and whether there are any problems with the generator-PC communication.

[Johnny B Good]

They've been using the current main board with its own on-board +/- 13.5 dc-dc converters for, afaicr, just over six months now, initially only using the earlier psu's 5V output with the +/- 13v rail components removed or their outputs disconnected. I see this new smaller psu board still has PTHs (unused) for the +/- 13v rails but no sign of the required components.

 Perhaps Feeltech are hedging their bets by not burning their boats over the possible need to revert to a three rail psu. However, in view of their track record on warranty repairs, I find this a little hard to believe. If you're going to the trouble to revamp the PCB layout into a more compact 5 volt only form, why include any sort of three rail provision?

 BTW, have you checked for continuity between the BNC grounds and the PE pin? It's an improvement in that they're using a dedicated wire (even if it doesn't conform to the colour coding for a safety earth wire) rather than pinching one of the DC ground returns to divert past the fuse to the PE tag but only if it's not relying on a DC ground trace running, it seems to me, perilously close to what appears to be a mains current inrush limiting PTC.

 If Feeltech have been lurking in these topic threads, it's just possible they may have taken inspiration from them to wire a 1 to 10 K resistor in series with earth connection.

[Labrat101]

Hi That PSU has to be the cheapest of cheap . The earth looks like it connected to the NTC  .. please tell me I am wrong . .
 Just love the new markings on the Cyclone .. Feelelec Pro 1941     Maybe your lucky its a Pro Fake 

[john_mallord]

BTW, have you checked for continuity between the BNC grounds and the PE pin? It's an improvement in that they're using a dedicated wire (even if it doesn't conform to the colour coding for a safety earth wire) rather than pinching one of the DC ground returns to divert past the fuse to the PE tag but only if it's not relying on a DC ground trace running, it seems to me, perilously close to what appears to be a mains current inrush limiting PTC.

 If Feeltech have been lurking in these topic threads, it's just possible they may have taken inspiration from them to wire a 1 to 10 K resistor in series with earth connection.

Yes, there's continuity between the BNC grounds in the front and back panel to the mains earth terminal.
I don't seen any resistor connected in series to the mains earth though, does it supposed to have a resistor in it?

What I found is that the DC ground (-) is connected to the mains earth via a thin PCB trace - pls see attached pic.
I'm not too sure if the two are supposed to be connected or should be isolated 

Hi That PSU has to be the cheapest of cheap . The earth looks like it connected to the NTC  .. please tell me I am wrong . .
 Just love the new markings on the Cyclone .. Feelelec Pro 1941     Maybe your lucky its a Pro Fake 

Yep, definitely cheap lol
The thermistor is connected to the Live terminal.

This whole thing weighs lighter than my phone and that FPGA marking is quite funny, does it mean that the cyclone chips were fake that they were forced to remove it?

I am interested in whether it is the FY6900-100M model or another, and whether there are any problems with the generator-PC communication.

Haven't tried the PC connection yet, still waiting for the 50 ohm feed-through that I ordered last week.
This is the FY6900-60mhz model from Bnggood btw.

[Labrat101]

The Cyclone chip in the FY6800 is definitely a FAKE (not even close) .
JohnnyBGood had the FY6600 and that worked far better and probably got an original . But as Cyclone is made under license from Intel from its daughter
company . The top of an original has a different surface like all Intel chips easy to spot .
For the power supply . Not really surprised keeping up their super low standards .   Also noticed that the YX820 6.3v bit close and will fail on a 5v.
The main board the only difference I can see from the 6800 is that they have just replaced all the voltage regs for a Buck inverter .
They still have not corrected the connections from Clown chip to the 2 DAC's . .  also noticed that there is an added chip under tcxo maybe 3N502
Clock Multiplier with a further effort to TRY and reduce the Jjjjiiitteerr  as the Fake chip does not have this feature built in ( as its a patented feature of Intel)
 .
 Buck for Buck they are not Bad if you are not expecting high end results .
We have done many upgrades to these units over the last few years . Its a real shame that FeelTeck never really took any of our ideas and applied them.
OK it would have increased the price by maybe $20 -$30  but in production they would have won .

[cdev]

Dear FeelTech tech support.

I have a FY 3200S version 2.0. I am largely happy with my Feeltech, with the exception of two problems. One is that it has the high voltage to ground, which means that I aways have to ground it using the BNC connectors. Which means I cannot just use it normally, without a good ground attached, or risk damaging chips. The second is that the calibration procedure does not work for me. As I was one of the early buyers of he 3224s.

Later versios of the firmware which I dont have appear to have an ability to calibrate the signal generation to an external reference signal. This would make it much more useful. My unit boots and 2.0 is shown on the LCD screen. Alternatively, is there a procedure to calibrate my function generator via an analog pot? (It been a while since I have opened it up now, I forget if it has a pot inside,if it does could you verify if that is what I woud use to adjust the frequency to a correct one. ) Alternatively, could it accept an external 10 MHz reference clock? If so, where would it be connected on the PCB? I can make the modification myself and add a BNC connector to use for the input.

Really ?!?
This is all what I get there:
模板不存在：/www/wwwroot/en.feeltech.net/template/default_html5/message/error.html 

My advice:  Flowing us with streams of words is counterproductive, believe me.
So, less writing, more reading, thinking. & checking...

Dear Customers,
I'm very apologized to I added an extra point after the link. 
This is my fault, sorry again.

From a mistaken employee。

Please dont worry about your English, its fine. Its the thought that counts.

[Johnny B Good]

@john_mallord

 Regarding the earthing, no surprises there. It's the same ol' same ol' hard earth contact between the noise polluted mains wiring with its random DC offsets and the BNC grounds as per the 6800 and previous versions of the 6900.

 All that's different is an improved routing via what looks to be a much better quality double sided psu PCB. It has more than adequate creepage distances and isolation slots between the DC output and all the high voltage mains connected stuff on the primary side of the transformer so I'd be less concerned over safety compared to the earlier PSU boards (in my case, a tall order since I had little enough safety concerns with the earlier psu board).

 I noticed they've used a 5A 60V PIV rated smd Schottky barrier rectifier diode (SS56). Not quite the best choice - a more optimal choice would have been an SS54 (5A 40V piv) for its lower forward volt drop to improve efficiency on a 2A rated 5 volt supply. The 6.3v rated smoothing cap is more than adequate since voltage peaks are unlikely to exceed the 5v output by more than 100mV or so.

 The thin grounding trace appears to still be visible were it goes over the chunk of ground plane by the DC output end. Checking the earlier pictures, this seems to have been doubled up on the top side. Is that actually the case?

 I was going to suggest that such a thin trace could act as a safety fuse in the event of a full live contact from a DUT being applied to the BNC grounds but if that trace has been doubled up, that sort of protection is compromised (ie. less likely to save the black earth wire from being burned to a crisp under such a fault condition).

 However, if you insert a 1 to 10 K resistor in series with the earth wire to keep mains noise and random DC offsets at bay, it can act as the safety fuse to guard against this fire risk (provided you mount the resistor clear of anything flammable).

[Vedius]

Hello, friends! For Christmas, I gave myself a KKMOON FY9900 60M. However, my joy was short-lived! If at first there were no problems with the operation of this device, now I see terrible things on the oscilloscope screen: a sinusoid curved inside out from below (in Sine mode) and squares full of noise on top (in Square mode). The rest of the charts are just as awful. The restoration of the waveform is possible only with the potentiometers W4 and W16 (CH1 and CH2) completely turned out, but the signal level becomes indecently small. I don't know what to do! The autopsy of the patient showed that I became the owner of a modified version V2.31 with built-in power +- 13.6V and with a counterfeit FPGA. The seller does not respond. Does it make sense to write in support of FeelElec? I count on the help of experienced users.

[john_mallord]

The thin grounding trace appears to still be visible were it goes over the chunk of ground plane by the DC output end. Checking the earlier pictures, this seems to have been doubled up on the top side. Is that actually the case?

I was going to suggest that such a thin trace could act as a safety fuse in the event of a full live contact from a DUT being applied to the BNC grounds but if that trace has been doubled up, that sort of protection is compromised (ie. less likely to save the black earth wire from being burned to a crisp under such a fault condition).

However, if you insert a 1 to 10 K resistor in series with the earth wire to keep mains noise and random DC offsets at bay, it can act as the safety fuse to guard against this fire risk (provided you mount the resistor clear of anything flammable).

Yes, there is also a thin trace, prolly around 4~5mil coming from the dc ground vias going to the top side of the pcb.


I'm back reading to check and see if someone has made a mod of using a battery pack to replace the psu?
Does it makes sense to do it regarding safety and to workaround the ground loops? Just got my SDS1104x-e last month so I'm very new on using this kind of test instruments.

Though I mostly work on battery powered and low signal microcontroller stuff (hobby), I'd knew that there would come a time that I have to work on some AC HV devices too.

[Johnny B Good]

@john_mallord,

 I was going to suggest placing the resistor on the PSU board to bridge a gap cut out of the top trace. The doubled up grounding trace means you'd also have to create a gap directly underneath in the bottom trace before drilling a couple of 0.6mm holes for a 1/4W 1 to 10 k wire ended resistor, ideally stood off from the board by 3 to 4 mm with ceramic beads.

 In view of this extra complication (and the more permanent modification of the PSU board), a better, reversible option would be to place the resistor in series with that black earthing wire. That way, if felt necessary, you can at least undo the modification without leaving any obvious signs of it ever being attempted (unsolder the wire from the PE tag to insert the resistor into circuit).

 As to the question of powering it from a battery, a good starting point would be Bad_driver's earlier contribution to this very topic thread right here:

https://www.eevblog.com/forum/testgear/re-feelelec-new-arrival-fy-6900-signal-generator/msg3836159/#msg3836159

 Since you only need a single 5v 2A max supply, an alternative option could be to use one of the later versions of battery banks now available which can act as a mini 5 volt UPS allowing you to wire it up between the existing 5v psu and the mainboard via a separate 2 pole change-over switch to save discharging the battery when not being used.

 This would allow you to disconnect its output and allow the existing PSU to power it directly as before whilst still providing charging current to the power bank. However in this case, I think you may need to upgrade the PSU to a 3 amp rated one to avoid overloading the original in a worst case scenario of low battery state charging demand (about an amp) on top of a worst case demand from the main board (possibly a maximum close to the 2A limit).

 An alternative to this arrangement would be to use a separate usb socket for charging the power bank and use a single pole change-over switch to select between battery or mains power operation, allowing you to retain the original psu board.

 Since the psu board is a class II double insulated unit, no safety earth connection is required. For our purpose we really only need access to a PE tag to connect a relatively low resistance "Static Drain" resistor to attenuate the half live mains voltage down to a vanishingly small value (0.25vac or less with the suggested resistor values). Even a drain resistor value as low as 1K will attenuate the unwanted noise and random DC offsets, polluting the mains earth wiring, by some 60dB or so.

 'Scopes, otoh, tend to use class I psus which demand an actual protective earth connection and it's generally considered not a good idea to disconnect this protective earth in order to eliminate any such ground loop issues, even when this is only a "temporary measure".

[Vedius]

Hello, friends! Without waiting for an answer to my question, I started looking for a problem. The fog has cleared up a bit! In my version of the main board, instead of two separate MCP4822 E operational amplifiers, the LMV324I is installed. In an unsuccessful attempt to find the culprit in the distortion of the waveform, I touched the tip of the soldering iron to the first output of the LMV324I. And, lo and behold! the waveform has recovered! True, after the generator was turned off, it returned to its previous state, but confidence in the success of the case returned to me. Now I have new questions:
1. Is the LMV324I working or faulty?
2. what leads to the fact that the LMV324I contributes to the distortion of the signal?
3. why is the signal restored when touching any of its pins (not just the first one - I checked)?
4.   why are there places near his conclusions without installed parts?
5. why do all the installed parts have standard denominations that are multiples of ten?
6. how to make the circuit work? (maybe you should change the LMV324I? - and if the problem is not in him?)
Someone, please respond!

[Labrat101]

As the picture is not so clear LMV324I may not be a Ti original . I would take a closer look . LMV324i is a pretty cheap dual rail op amp
as Dave would say its a Jellybean . I would replace it . most of the final stage was a mess . in the 68 version as well .

[Labrat101]

Here is the Diagram of the 66/68 version the power supply is different but the rest looks like its about the same .
May help you .

[Bad_Driver]

After changing the PSU and the XO against better ones and than go back to the state of origin I have now a full battery driven generator.
See my earlier posts. (6x 18650). This was the best improvement of all.

And now I use it much more. Without the stiff IEC cable and a running time of +6h it became a handy device.
But for serious work I have the Siglent SDG as well. I'm now looking for battery driven scope as supplemental to my big one.

What about a OWON XDA3062X??

[Vedius]

So, I didn't wait for practical advice and decided to reanimate my FY6900 myself. In my version of the main board, instead of two MCP4822e ADCs, an LMV324i quad op-amp is installed. My initial suspicions about its malfunction were not justified. Therefore, I replaced JRC4556, which did not inspire confidence in me, with OPA2132: the sine wave was restored, but it had a strong upward bend from below. The adjustments W5/W26 and W3/W16 were not enough to correct the bend, so I added a gain at the output of OPA2132, for which I replaced the resistors R57 and R58 with 3K to 10K, as well as the obviously low R59 and R60 with 10Ohm to 100Ohm. After that, I calibrated the offset to OPA2132 and adjusted the gain of the AD8009 on two channels. Previously, I changed the resistors at the inputs, corrected the resistance of the output attenuators from 85 ohms to 50 ohms, made a linear power supply for three voltage values, turned off the built-in voltage converter +13.5V /-13.5V, installed a 4cm 5V fan аnd shielded the main board with a metal grid. Now my generator is working properly. Sorry for my clumsy English 

[Henry Finley]

So here I am on a Sunday morning about to pull the trigger on a Dominty FY6900 from Amazon after several days of research on these Feeltec 68 and 6900's. And as a last minute click on this thread that caught my eye, it seems the rug was pulled out. I read a lot of this. And much of it looked good. Now I'm wondering if it isn't just the worthless Chinese crap I originally suspected, and that I'd be throwing my money away. All I wanted to do was have a little accuracy aligning my tape recorders, vintage shortwaves and vintage test equipment. I'm not running Los Alamos nuclear research over here. I figured it would likely suit me. Now its begining to look like I'd be better off peeling off $125 in cash and walking to the dumpster and throw it in.

[radiolistener]

now I see terrible things on the oscilloscope screen: a sinusoid curved inside out from below (in Sine mode) and squares full of noise on top (in Square mode). The rest of the charts are just as awful.

it looks like issue of your oscilloscope.

Can you show results taken on more serious oscilloscope?

[Labrat101]

So here I am on a Sunday morning about to pull the trigger on a Dominty FY6900 from Amazon after several days of research on these Feeltec 68 and 6900's. And as a last minute click on this thread that caught my eye, it seems the rug was pulled out. I read a lot of this. And much of it looked good. Now I'm wondering if it isn't just the worthless Chinese crap I originally suspected, and that I'd be throwing my money away. All I wanted to do was have a little accuracy aligning my tape recorders, vintage shortwaves and vintage test equipment. I'm not running Los Alamos nuclear research over here. I figured it would likely suit me. Now its begining to look like I'd be better off peeling off $125 in cash and walking to the dumpster and throw it in.

Just forget FY it stands for Friging Yunk . the chip set is 99.99% fake clones and then every one wonders why the sin is funny and the square looks shaky
Go for a UNI-T 962 they are about the same money and they work fine out the box . They may not be up-to NASA specs . But dam close .
There is no jitter and it uses good qual chips .   

[Henry Finley]

So here I am on a Sunday morning about to pull the trigger on a Dominty FY6900 from Amazon after several days of research on these Feeltec 68 and 6900's. And as a last minute click on this thread that caught my eye, it seems the rug was pulled out. I read a lot of this. And much of it looked good. Now I'm wondering if it isn't just the worthless Chinese crap I originally suspected, and that I'd be throwing my money away. All I wanted to do was have a little accuracy aligning my tape recorders, vintage shortwaves and vintage test equipment. I'm not running Los Alamos nuclear research over here. I figured it would likely suit me. Now its begining to look like I'd be better off peeling off $125 in cash and walking to the dumpster and throw it in.

Just forget FY it stands for Friging Yunk . the chip set is 99.99% fake clones and then every one wonders why the sin is funny and the square looks shaky
Go for a UNI-T 962 they are about the same money and they work fine out the box . They may not be up-to NASA specs . But dam close .
There is no jitter and it uses good qual chips .

Thank you. I will check it out, but wth due suspicion over anything coming from China. I have a basic resentment of sending a penny there. I have also checked out an HP 8111a from the 1980's that I can get in unknown condition (but actually reparable). I'm becoming research-weary.

[Labrat101]

So here I am on a Sunday morning about to pull the trigger on a Dominty FY6900 from Amazon after several days of research on these Feeltec 68 and 6900's. And as a last minute click on this thread that caught my eye, it seems the rug was pulled out. I read a lot of this. And much of it looked good. Now I'm wondering if it isn't just the worthless Chinese crap I originally suspected, and that I'd be throwing my money away. All I wanted to do was have a little accuracy aligning my tape recorders, vintage shortwaves and vintage test equipment. I'm not running Los Alamos nuclear research over here. I figured it would likely suit me. Now its begining to look like I'd be better off peeling off $125 in cash and walking to the dumpster and throw it in.

Thank you. I will check it out, but wth due suspicion over anything coming from China. I have a basic resentment of sending a penny there. I have also checked out an HP 8111a from the 1980's that I can get in unknown condition (but actually reparable). I'm becoming research-weary.

Your in the USA there are a lot of second hand stuff I see on Ebay all maybe 20 years old . Get an old HP signal gen or something else the old school
stuff can be repaired . most of it is still working better than the new crap. My whole lab is old school except the uni-t which is about a year .
there are UNT-T agents in the US . no need for china .

[nctnico]

So here I am on a Sunday morning about to pull the trigger on a Dominty FY6900 from Amazon after several days of research on these Feeltec 68 and 6900's. And as a last minute click on this thread that caught my eye, it seems the rug was pulled out. I read a lot of this. And much of it looked good. Now I'm wondering if it isn't just the worthless Chinese crap I originally suspected, and that I'd be throwing my money away. All I wanted to do was have a little accuracy aligning my tape recorders, vintage shortwaves and vintage test equipment. I'm not running Los Alamos nuclear research over here. I figured it would likely suit me. Now its begining to look like I'd be better off peeling off $125 in cash and walking to the dumpster and throw it in.

Just forget FY it stands for Friging Yunk . the chip set is 99.99% fake clones and then every one wonders why the sin is funny and the square looks shaky
Go for a UNI-T 962 they are about the same money and they work fine out the box . They may not be up-to NASA specs . But dam close .
There is no jitter and it uses good qual chips .

Thank you. I will check it out, but wth due suspicion over anything coming from China. I have a basic resentment of sending a penny there. I have also checked out an HP 8111a from the 1980's that I can get in unknown condition (but actually reparable). I'm becoming research-weary.

Whatever you buy new gets made in China anyway nowadays and older generators (from the 80's or 90's) are in much worse shape compared to what you can buy from Uni-T or Feeltech. Old test gear is highly overrated (been there, done that). If I needed a simple but versatile function generator nowadays, then my choice would be the UNI-T 962 or the FY6900 (and not mess with it). These runs circles around the HP 8111A where it comes to frequency stability and ease of use.

[Henry Finley]

So... assuming this is the one I have settled on, what is this "pre-order" bit on the Banggood site? Pre-orders arrive on Feb 10 it says. And this is Jan 30. Does that mean they have none? Or does it mean if they did, they're not even going to bother to ship till AT LEAST Feb 10, if not later. Will I get it before the rapture or the zombie apocalypse?