EEVblog #978 - Keysight 1000X Hacking

[candrian]

Do you know if i remove the screws and the back cover will void the warranty?

[JDubU]

I don't know.

[TK]

Do you know if i remove the screws and the back cover will void the warranty?

In the US it does not void the warranty.  Every country has different warranty laws, so you should ask your local Keysight rep.

[candrian]

As it's getting really disturbing while working i removed the back cover to check if the noise changes. But the result is the same. The noise is coming from the fan itself.

Check the video.

https://drive.google.com/file/d/13oBXFKfoMoiRLk5rrGS71VSptFmQdDiC/view?usp=sharing

I wouldn't expect this noise from an instrument that costs a thousand

[JDubU]

In addition to the fan grill modification, I replaced the fan in my first generation 1000x with a Noctua NF-A6x25 FLX. 
Don't know if it's appropriate for the second generation 1200x.

[TK]

As it's getting really disturbing while working i removed the back cover to check if the noise changes. But the result is the same. The noise is coming from the fan itself.

Check the video.

https://drive.google.com/file/d/13oBXFKfoMoiRLk5rrGS71VSptFmQdDiC/view?usp=sharing

I wouldn't expect this noise from an instrument that costs a thousand

It must have a faulty fan.  Mine had lower noise than the first generation 1000X

[candrian]

I can confirm DSOX1202 has the same FAN as 1102

HengLiXin HD6020B12M

[candrian]

I finally replaced the FAN with a Noctua NF-A6X25 and everything now sounds a lot smoother.

I created a small video in case anyone has the same issue and wants to hear samples
https://youtu.be/ekAmelIheRk

[rbaleiro]

Hi guys,

Here is my history of trying to modify the EDUX1002A into DSOX1002A.

I bought all the components (a few imported and others directly from here, Brazil). When unmounting the parts one of the white plastic connectors in the logic board became loose in the board and I've tried to put it back with hot air... result: it melted.

A long time after trying to realize the type of connector and brand, finally found it. Luckily got it as a sample and soldered it. (https://www.te.com/usa-en/product-5177984-1.html)

All the other components were put in place as specified. Particularly the WSON LMH6552SD is hard to tell if its in place correctly... because the contacts are bellow the IC. That may be one of the issues I am facing.

When i finally put that to work... the oscilloscope turned on, but any signal I put in the channels 1 or 2 doesnt show anything. Neither the test function is working... it is acting like there was no probe. May be the trigger.. may be the WSON IC... how to know?!

Another thing is the product ID. Even replacing the resistor it still shows EDUX in the screen. Is there anything I am missing? My firmware version is FERCSA's 1.10.

Happy to hear some tips on how to solve it.

update:  updated the firmware to 1.20 and still working the same.

[TK]

Another thing is the product ID. Even replacing the resistor it still shows EDUX in the screen. Is there anything I am missing? My firmware version is FERCSA's 1.10.

The model number in the screen will never change, as it is programmed somewhere protected in the FLASH memory, the same for the serial number

[TK]

Hi guys,

Here is my history of trying to modify the EDUX1002A into DSOX1002A.

I bought all the components (a few imported and others directly from here, Brazil). When unmounting the parts one of the white plastic connectors in the logic board became loose in the board and I've tried to put it back with hot air... result: it melted.

A long time after trying to realize the type of connector and brand, finally found it. Luckily got it as a sample and soldered it. (https://www.te.com/usa-en/product-5177984-1.html)

All the other components were put in place as specified. Particularly the WSON LMH6552SD is hard to tell if its in place correctly... because the contacts are bellow the IC. That may be one of the issues I am facing.

When i finally put that to work... the oscilloscope turned on, but any signal I put in the channels 1 or 2 doesnt show anything. Neither the test function is working... it is acting like there was no probe. May be the trigger.. may be the WSON IC... how to know?!

Another thing is the product ID. Even replacing the resistor it still shows EDUX in the screen. Is there anything I am missing? My firmware version is FERCSA's 1.10.

Happy to hear some tips on how to solve it.

If the connector that you resoldered is the one to the left, that is where the input signals are going to the ADC, so I suggest you to check the soldering of the connector one more time.  It is strange that both channels are gone, so I suspect it can be this connector.  Or it can be the WSON IC, as you said, it is very hard to solder.  Remove them and check for bridges on the pads, clean and resolder with very few solder

[rbaleiro]

Hi guys,

Here is my history of trying to modify the EDUX1002A into DSOX1002A.

I bought all the components (a few imported and others directly from here, Brazil). When unmounting the parts one of the white plastic connectors in the logic board became loose in the board and I've tried to put it back with hot air... result: it melted.

A long time after trying to realize the type of connector and brand, finally found it. Luckily got it as a sample and soldered it. (https://www.te.com/usa-en/product-5177984-1.html)

All the other components were put in place as specified. Particularly the WSON LMH6552SD is hard to tell if its in place correctly... because the contacts are bellow the IC. That may be one of the issues I am facing.

When i finally put that to work... the oscilloscope turned on, but any signal I put in the channels 1 or 2 doesnt show anything. Neither the test function is working... it is acting like there was no probe. May be the trigger.. may be the WSON IC... how to know?!

Another thing is the product ID. Even replacing the resistor it still shows EDUX in the screen. Is there anything I am missing? My firmware version is FERCSA's 1.10.

Happy to hear some tips on how to solve it.

If the connector that you resoldered is the one to the left, that is where the input signals are going to the ADC, so I suggest you to check the soldering of the connector one more time.  It is strange that both channels are gone, so I suspect it can be this connector.  Or it can be the WSON IC, as you said, it is very hard to solder.  Remove them and check for bridges on the pads, clean and resolder with very few solder

The connector that broke was the one near the Product ID. That may be the case it is not recognizing the modification.... still, the conector is perfectly fit.

Below are some pictures of the state of the oscilloscope. The trigger function has always to be in normal mode to show anything... cause it isnt synchronizing. The square shaped curves are measured in the test point without any function. The curvy shapes are the sine wave being generated by the HELP->sine generation. The signals not synchronized is just to show how it stays.

Please, if you can explain how to properly make the WSON IC soldering... cause I saw several tutorials on youtube and still have my doubts about it. If u can be clear about amount of tin, flux, air pump pressure and temperature I would really appreciate!

[JDubU]

This is what works for me:

Use:
1.  63/37 (eutectic) 0.02" diameter leaded solder (like Kester 44).
2.  A soldering iron tip that is chisel shaped and as close to the width of the pad as possible.  I use a 700 degree (F) tip temperature for soldering and an 800 degree (F) tip temperature for desoldering.
3.  Copious amounts of liquid flux!  I use a Kester 186 flux pen.  Amtech no-clean tacky flux is highly regarded but I have yet to try it.
4. 99+% isopropyl alcohol with a brush to remove flux.

For really small SMD parts (like the WSON package):
1. For rework, remove all lead-free solder from the pads using flux and copper braid.
2. Tin pads with leaded solder and then remove excess with flux and copper braid, so pads are perfectly flat.
3. Coat pads with plenty of liquid flux.
4. Place part on fluxed pads and hold in place with fine tweezers.
5. Wipe soldering iron tip on wet sponge to remove any solder or burnt flux.  Tip needs to be perfectly clean and shiny.
6. Melt a tiny amount of solder on the very end of the soldering iron tip.  This is the only solder that will be used to make the solder joint!
7. Briefly touch the solder that is on the end of the tip to the junction between the part and the pad and then smoothly pull the tip away along the surface of the pad.  Surface tension of the solder will cause excess solder to be removed from the joint.  The flux will make sure that the solder flows properly into the joint. 
8. Clean off flux with isopropyl alcohol.

[rbaleiro]

This is what works for me:

Use:
1.  63/37 (eutectic) 0.02" diameter leaded solder (like Kester 44).
2.  A soldering iron tip that is chisel shaped and as close to the width of the pad as possible.  I use a 700 degree (F) tip temperature for soldering and an 800 degree (F) tip temperature for desoldering.
3.  Copious amounts of liquid flux!  I use a Kester 186 flux pen.  Amtech no-clean tacky flux is highly regarded but I have yet to try it.
4. 99+% isopropyl alcohol with a brush to remove flux.

For really small SMD parts (like the WSON package):
1. For rework, remove all lead-free solder from the pads using flux and copper braid.
2. Tin pads with leaded solder and then remove excess with flux and copper braid, so pads are perfectly flat.
3. Coat pads with plenty of liquid flux.
4. Place part on fluxed pads and hold in place with fine tweezers.
5. Wipe soldering iron tip on wet sponge to remove any solder or burnt flux.  Tip needs to be perfectly clean and shiny.
6. Melt a tiny amount of solder on the very end of the soldering iron tip.  This is the only solder that will be used to make the solder joint!
7. Briefly touch the solder that is on the end of the tip to the junction between the part and the pad and then smoothly pull the tip away along the surface of the pad.  Surface tension of the solder will cause excess solder to be removed from the joint.  The flux will make sure that the solder flows properly into the joint. 
8. Clean off flux with isopropyl alcohol.

Thank you for the reply. Unfortunately this WSON package doesnt have a terminal on the side, just at the bottom. No way I can use the iron, has to be with hot air.

[JDubU]

Thank you for the reply. Unfortunately this WSON package doesnt have a terminal on the side, just at the bottom. No way I can use the iron, has to be with hot air.

I have used that technique successfully on those LMH6552SD WSON packages.  The key elements are pre-tinning the pads with leaded solder, lots of flux, and the precise alignment of the part on the pads.  The liquid flux is essential to getting the solder to smoothly flow into the junction between the terminals underneath the part and the pad.

[rbaleiro]

Thank you for the reply. Unfortunately this WSON package doesnt have a terminal on the side, just at the bottom. No way I can use the iron, has to be with hot air.

I have used that technique successfully on those LMH6652SD WSON packages.  The key elements are pre-tinning the pads with leaded solder, lots of flux, and the precise alignment of the part on the pads.  The liquid flux is essential to getting the solder to smoothly flow into the junction between the terminals underneath the part and the pad.

Ive checked the terminals and apparently there is a tiny terminal in the side of the package. I have to buy a more precise tip, but i couldnt find a chisel type that is too small as the terminal (0.25mm). Just cilindrical/conical. Is it a problem? The only flux i have is this paste like the "no clean flux" you said... but i have no idea if that is no clean. What do u think?

[rbaleiro]

Another thing is the product ID. Even replacing the resistor it still shows EDUX in the screen. Is there anything I am missing? My firmware version is FERCSA's 1.10.

The model number in the screen will never change, as it is programmed somewhere protected in the FLASH memory, the same for the serial number

So what is the point of replacing the Product ID thru the resistors, tho? How can I check if the change is working?

[TK]

Another thing is the product ID. Even replacing the resistor it still shows EDUX in the screen. Is there anything I am missing? My firmware version is FERCSA's 1.10.

The model number in the screen will never change, as it is programmed somewhere protected in the FLASH memory, the same for the serial number

So what is the point of replacing the Product ID thru the resistors, tho? How can I check if the change is working?

The information you see in the screen is not changed, but the scope will boot into the mode set by the resistors.  You will see features that are only enabled on the DSOX, even when the information says EDUX...
The HW mod is required as the EDUX comes limited to 50MHz and no ext digital input on the additional BNC.  By adding the missing components and changing those that are different, you are enabling the HW to be ready to the soft configuration that is liberated by the FERCSA Hack.

[JDubU]

I used a 0.031" (0.79mm) chisel tip, turned so the width of it was vertical.  Chisel tips are much preferred for this since they can feed solder to the joint more easily than a conical tip.  Due to surface tension, a conical tip will tend to form a ball of solder at the tip which then places a blob of solder on the work.  A chisel tip has a small reservoir of solder on the face of the chisel which can feed solder as needed by the size of the joint.
I highly recommend using liquid flux for small SMD parts.  Paste flux can be too thick, causing the part to float and move as the paste begins to melt.

For me, precisely positioning the WSON over the thin pads was the most challenging.  First time I tried it, the terminals centered themselves between adjacent pads.

[rbaleiro]

Finally it worked! What made the real difference was using a 0.2mm copper soldering tip B900M-TI and the NO CLEAN FLUX. In addition I had a problematic resistor in the channel 1 causing me headaches.

Now I realized that the measurement is slightly out of calibration. I have already done the automatic user calibration from the scope and the probe capacitance calibration.

Is there any way we can calibrate it to make it more precise? I feel the trace is too thick and the peak voltage usually is higher than it should be.

[GN_ghost]

Try this link to a full package (~30Mb) with a backup of the original 1.20 in Image2:

https://we.tl/t-GW2QZ3NnjQ

Hi Bud,

I am trying to tryout your package on my DSOX1102G. It seems that the link has dead. Can I ask if we could get it refreshed? Thanks~

[Bud]

Grab it from this post:

https://www.eevblog.com/forum/testgear/dsox2000-and-3000-series-licence-have-anyone-tried-to-hack-that-scope/msg3388792/#msg3388792

User PhillyFlyers kindly provided a storage for it.   

[GN_ghost]

Grab it from this post:

https://www.eevblog.com/forum/testgear/dsox2000-and-3000-series-licence-have-anyone-tried-to-hack-that-scope/msg3388792/#msg3388792

User PhillyFlyers kindly provided a storage for it.   

Thanks a lot! I have tried upgrading from 1.10 liberated firmware directly. So far so good~

[GN_ghost]

I finally replaced the FAN with a Noctua NF-A6X25 and everything now sounds a lot smoother.

I created a small video in case anyone has the same issue and wants to hear samples
https://youtu.be/ekAmelIheRk

Hi candrian

I also have the original fan replaced by the Noctua one. Indeed part of the noise is gone, but there is still a high freq tone, not as good as you shown in the video. Not sure if the fan grill mod is also necessary...

[SilverSolder]

The most important mod is a rubber suspension for the fan, so motor vibrations don't transmit into the "sound board" of the equipment enclosure...   Often this is enough to make the stock fan tolerable.