FeelTech FY6600 60MHz 2-Ch VCO Function Arbitrary Waveform Signal Generator

[cybermaus]

Sorry to hear that. Unfortunately you seem to be the 5th one where the device went bad.

It may be better to move your question to the other thread though, so we keep the flash corruption & problem emails in one place.

At the very least, read that thread, because it seems your failure mode is completely different then the other 4, who had all 4 had identical progressive problems, and yours is a sudden dead. So it may simply be a power problem on the display board.
Hopefully so, because I have a STLink2 flash/read tool underway, but it seems often these STM32 chips are read-protected in commercial products.

Also: So flashchip is completely identical. Ok. What version and frequency model?

[chtech0710]

Sorry, my fault. I just googled "FY6600 firmware" and as a result found this post.
Now i've read the other thread, but it doesn't help very much.

I don't think it's a power problem, because the 3.3V on the vreg for the STM are present and stable. (Perhaps, a transient killed the STM32? Mhmmm...)
I've tried to measure if the crystal (next to the STM32) is working, but nothing. I replaced it, nothing changed.

I've got the 30 MHz model. I think the version was the latest, i can't tell you exactly, cause i'm at home now and not at my workshop.

The content of the flash chip ist almost identically (havent checked every single bit  )
I've read out the old firmware from the flash chip (and saved it). Have tried your uploaded one. Makes no difference if I try your firmware or the original one.
I do not think the Cyclone is broken. By capturing some data between the flash chip and the Cyclone, everything seems normal.

BTW, could a friendly moderator move my (and your last) post into the appropriate thread? THX 

[Diabolo]

Hello,

To read, write, remove read / write protection on STM32:
- http://www.jyetech.com/Support/STM32_FlashLoaderDemonstrator_v2.7.zip

Diabolo

[chtech0710]

Just recently used this tool to upgrade the firmware from my DSOShell, but I wouldn't have come up with the idea to use it for this problem. 
But, this tool needs a bootloader to work? i dunno.

Anyhow, i will try it tomorrow. THX 

[cybermaus]

From what I read in this STM32 doc (yes, I have been reading a lot about the STM32 lately. Demo board and STLink2 tool should be here tomorrow), the bootloader is always present on this chips, in masked ROM even. Hence it can update using this serial tool.

But remember, you want to read, not write, the firmware.

But good to check the tool anyway, at least it will tell you if the chip is still alive.

Can you send me your WInbond flash file, so I can binary compare them?
ALso, I would really like to know what version your device was. Did you never look before it crashed?

[chtech0710]

Unfortunately, I do not have much experience with STM32 µCs. At work we use mostly ATmega, PIC, some lower-end FPGA or "fancy" embedded computers with linux. (it's a small company)

Good to know that the bootloader is always present. I will definitely try to unlock and/or read the content from the STM32 tomorrow (but only if the chip is alive  )
And i will also upload the flash content from the Winbond flash ic.

Apart from that, by comparing available pictures of the hardware it almost looks like the hardware is pretty identical to most models (15/30/50/60?MHz). A comparison of the different firmwares should give us information about possibilities for hacks. 

I was recently looking for an update of the firmware, but it was (and is) not possible yet, and unfortunately i didn't take much care about my fw version. I#m pretty sure it was the latest version 3.1(?)

[cybermaus]

Careful with the "unlock". Not sure if you can do that from serial, but:

According to these people, unlock from security level RDP1 to RDP0 is possible, but will auto-erase the flash, making the chip back ready for a new program, but not helpful for us.

They did find a way to read the flash while in RDP1, by means of a custom reader that resets the debug tool, it seems that each well timed reset leaks one byte of flash. But that will not work for normal available tools, and they did not release the software.

From what I understand, our only realistic hope is that the manufacturer did not secure the chip. Or makes the firmware available in order to stop negative posts.

[chtech0710]

After many cups of coffee, and playing around i've managed to get access to the bootloader with the ST Flash Loader Demonstrator.

First of all, the memory is protected... 
and by trying to unlock, it was auto-erased... damn!

In case someone else will try it:
USART1 isn't easy availlable, so you have to solder some small wires to the STM. The RX/TX from J1 are wired to USART2 and useless in our case.
First, we need to set the BOOT0 pin to high (3.3V) to enable the bootloader. BOOT0 is at pin 44.
USART1-TX is on pin 30 and USART-RX is on pin 31, which also goes to C1 and the encoder.
You have to remove C1, cause C1 acts like a filter. I've used the pad from C1 to attach the RX-wire.
We need the RESET-pin, which is availlable at pin 7 and R1.
GND and 3.3V are availlabe from JK1.

Finally, the baudrate is 115200,8N1 with even parity.
By using the Flash Loader Demonstrator, you shoud disable ECHO and set the timeout to 5.

And here are some pics showing my attempts to get access to the bootloader.

[cybermaus]

Big bummer. No need for me to check anymore if the STM32 is locked

So there are already 5 people with broken STM32 flash. And we know how to write the flash.
We need some way to make the vendor relinquish the firmware?

Easy for me to say, mine is still working, but keep asking for service requests? Not buy anymore of their stuff until this is fixed / fixable.
That latter one is pretty understandable, as much as I am happy with mine, knowing it can corrupt its own flash at any moment means I would not recommend to others.


Sidenote, I checked your firmware file. While it appears the same on a quick glance, it is quite different.
Not only did you already upload the first and last few arbitrary waveforms, 2 of the build in ones are also slightly different.
And though not easy to spot due to the low bit density, the Cyclone part is also different all over.
So it seems they are making changes to the Cyclone between versions as well as the UI

[beanflying]

Disappointing about the firmware failures. Time for Feeltech to release a method and the firmware to fix this issue and enable user upgrading or repair. 

This thread until it happens will continue to turn up very high in the search rankings and random failures that require unit returns to mainland China are not going to reflect well on sales.

[Scratch.HTF]

Since I intend to sync this unit to an external frequency reference, is the frequency at the Sync OUT port 10 MHz?

[Diabolo]

Hello,

Aie, it's a shame for the erased STM32 memory.
But we will find a solution to restore everything in condition.
----
How does a power supply / charger for laptops, LCD TVs, GSM, etc ... (Sony, Asus, Samsung, etc ...) that have a 110V-220V with 2 pins compared to the ground?
They must also have 32V of differential with the ground under power supply AC 220V and the general mass ?
----
Datasheet on various TCXO, stability seems good for general purpose :
- http://www.metatech.com.hk/product/fordahl/pdf/2002%20TCXO%20Page%2043-58.pdf
- http://docs-europe.electrocomponents.com/webdocs/0d5a/0900766b80d5a4cf.pdf


Diabolo

[cybermaus]

Since I intend to sync this unit to an external frequency reference, is the frequency at the Sync OUT port 10 MHz?

No, but its fairly easily and effectively added.

[beanflying]

Hello,

How does a power supply / charger for laptops, LCD TVs, GSM, etc ... (Sony, Asus, Samsung, etc ...) that have a 110V-220V with 2 pins compared to the ground?
They must also have 32V of differential with the ground under power supply AC 220V and the general mass ?

Diabolo

Some are good some are junk 

Quick test on what I have on hand  on the bench at present 2-40V above ground. Includes built in USB mains adapters. 17V dc LED drivers. The one that is the best is also the cheapest 

I have a bunch of these all are in the 2V ac range  https://www.ebay.com.au/itm/AC-DC-220-to-5V-12V-5V-600mA-Isolated-Switching-Power-Supply-Module-Power-Supply/400914188250?hash=item5d585907da:g:AYkAAOSwv0tVRd-R

[cybermaus]

Test the illogically good cheap ones again while reversing their AC polarity.

[beanflying]

Fairly humid and hot here tonight. To earth in bare feet on concrete I am generating about 200mVac  Also drinking my 3rd Beer and hot swapping 240AC what could go wrong 

column 1 is one way, 2 is reversed ac. First 4 including the 12V are the same manufacturer. Last 3 are just a few unearthed randoms LED drivers in the box, didn't bother reversing inputs need more

5Vdc 1 - 1.5Vac, 1.47Vac  5.0980Vdc
5Vdc 2 - 1.3Vac, 1.32Vac, 5.0860Vdc
5Vdc 3 - 1.46Vac, 1.45Vac, 5.084Vdc
12Vdc  - 1.14Vac, 1.13Vac, 11.992Vdc

3.16V ac 23.1Vdc
2.89V ac 20.9Vdc
75.1 Vac 12.1dc largest at 10Wand the most EXPENSIVE of the bunch 

So the conclusion from this is 

[aylesj]

If anyone was thinking of buying one, fasttech are offering 15% off everything at the moment (valid 23rd - 27th Nov), so the 60 MHz version can be had for $86.66 including delivery (Normally $101.95). Use the coupon code THANKS15 after you have added it to the basket.

Unfortunately I ordered one from them a week ago, so missed this discount. It took 8 days from ordering to receive an 'item shipped' email. I'm guessing it will take another 10-14 days to ship to the UK.

[plb]

Just received my FY6600. Free post, one week Shenzen to Sydney. Impressive.
This is my first post on this thread. I won't be hacking the instrument (much), but follow the thread with interest.

Not sure how to post a quote, but this from Cybermaus 3 weeks ago got my attention.
"So I took CH2, looped it into the VCO and that works but it is a bit unclean. In fact, I found it better to simply put a sawtooth on CH2 and do FM modulation, that also causes a frequency sweep. AM for amplitude sweep. For PWM, still have to use the CH2-VCO"
I find the Ch2 - VCO loop works well, but it is necessary to offset the Ch2 output otherwise you only get half a sweep. Ch2 swings around 0V, but VCO in expects 0-5V. Using CH2 with ramp in this way gives a trigger that is ideal for syncing the scope trace to the Feeltech.

Also I'm a bit intrigued by the power supply issue. As soon as I connect the Feeltech to a scope or other grounded instrument, its internal common is solidly grounded. Nevertheless I will be looking at a grounded IEC connector, as I do appreciate the risk of connecting the signal to a device under test before connecting the ground.

[beanflying]

Welcome 

My main concern was frying a low voltage micro in some way with the high floating ground. My bench has a dedicated earth point if needed regardless but I really wanted it to remain floating depending on the application but with lower risk to any DUT.

More risk now of me frying the DUT than the Feeltech 

[cybermaus]

Just received my FY6600. Free post, one week Shenzen to Sydney. Impressive.
I find the Ch2 - VCO loop works well, but it is necessary to offset the Ch2 output otherwise you only get half a sweep. Ch2 swings around 0V, but VCO in expects 0-5V. Using CH2 with ramp in this way gives a trigger that is ideal for syncing the scope trace to the Feeltech.

I have since created an dedicated ARB with a sawtooth from 0 to 1 rather then -1 to 1, but starting with a tiny -1 pulse to easily have the scope trigger.


Another usage tip I have since found (maybe common knowledge already)
People (including myself) complained both channels automatically start when you turn on the device. That is actually a setting in the system menu. (ChX boot Yes/No)

[Ebel0410]

I folks !

Finally I've decided not buying a new power supply, but making some modifications to the stock PS Feeltech.
I've focused my work on the caps and I would say that it's worth it.

The output filtering capacitors were swapped with low ESR and higher voltage ones.
Main front end circuit, a new high voltage cap 68µ/400v now takes place instead of the old crappy one (10µ/450v)
And last but not least, I've added a second HV cap between 0V and the unconnected mains.
It didn't cost me anything as I used parts from my old stock of electronic devices.


 

[beanflying]

I think your Power Supply board is now at 'capacity' 

Did you run an earth wire or just the extra Caps to try and soak up a bit of the voltage?

[Ebel0410]

I think your Power Supply board is now at 'capacity' 

Almost, but you know this PS is really a piece of sh..
I had to remove all the output diodes and replace them with more efficient ones. The gain is roughly +0,5V on each output.
Not so bad !

Did you run an earth wire or just the extra Caps to try and soak up a bit of the voltage?

Only an extra Cap so the leaked voltage is now half the value of the mains voltage.

[nctnico]

It didn't cost me anything as I used parts from my old stock of electronic devices.

So replace brand new components with old ones in unknown condition? 

[Specmaster]

I folks !

Finally I've decided not buying a new power supply, but making some modifications to the stock PS Feeltech.
I've focused my work on the caps and I would say that it's worth it.

The output filtering capacitors were swapped with low ESR and higher voltage ones.
Main front end circuit, a new high voltage cap 68µ/400v now takes place instead of the old crappy one (10µ/450v)
And last but not least, I've added a second HV cap between 0V and the unconnected mains.
It didn't cost me anything as I used parts from my old stock of electronic devices.

Why is that transformer at that ridiculous angle, surely you sorted that out?