Keysight 3000T scope 1 GHz(or more) hardware upgrade mod

[badbiki]

Lifted one pad on a removed relay... could patch if going backwards but otherwise alll relaysbin.

Kapton tape and insulation pads from  3D printer hot ends works well. Make wrap around a for the protecteded relays and a lifted one for the removal. Easy.

Not good enough to reuse but don’t damage the remaining relay

More tomorrow.

[badbiki]

MOd complete spent hours cleaning residue from all the traces and components.... so many hours!

I matched the 27.5 ohm resistors in pairs between 3 and 9mOhm difference. I would think it would be wise to also match the mirrored capacitors also, and match the mirrored inductors per channel.

Ideally I would have also bought more of C1 and C2 to match filter frequencies between channels. However it was late and I didn’t have a good 4wire measurement system setup for 0603/0805 passives. I’ll revise my BOM above and purhaps post my annotated drawings which print well on B&W and show movements/removals.

I’ll take it to my friends RF lab soon to check each channel at high frequencies. So far the user calibration passes so long as my cables T out to 1,3 and 2,4 otherwise it failed the first time.

Top hacking

[Samogon]

To measure 1pF caps is a big deal, not many systems could do it.

[badbiki]

I had a very quick hookup using a Rigol 70kHz-3.2GHz spectrum analyser set on 0dBm with each channel terminating into 50ohm with a 400mm BNC to SMA cable. Too quickly as I was rushing out the door to get home.

I will setup better for it next time to do a sweep up the frequency range and perform a segmented capture across the MHz and into the GHz.

Ch/Freq	1GHz (mVpp)	1.5GHz (mVpp)	1GHz (dBm)	1.5GHz (dBm)
Ch1	562.21	145.34	-1.02	-12.77
Ch2	491.96	125.38	-2.18	-14.06
Ch3	481.91	128.14	-2.36	-13.87
Ch4	511.31	20.653	-1.85	-29.72

I have a feeling that the last entry I captured was not setup correctly

I should have had this set as 3.97940 dBm for a clean 1Vpp signal.

I read a 1.699-1.718ns rise time on the trigger output with ~200MHz bandwidth measurement with Math FFT(d/d(t)).

[maxwell3e10]

Not exactly on topic, but I've been collecting shorted input noise spectra for various oscilloscope in this thread:https://www.eevblog.com/forum/testgear/oscilloscope-input-noise-comparison/

So far R&S RTM3000 is the winner, so to be fair we need some data from a Keysight 3000 series scope. It would also be interesting to compare the noise for mod and non-mod versions.

[badbiki]

Ye sure why not. That’s a lot of data to collect every div every channel

[1359944032]

I am going to buy 3014t and upgrade it to 1Ghz, but I don't know how to crack the license. Can someone tell me, thank you very much.
I am currently learning English

[badbiki]

Moving the jumpers “unlocks” the higher frequencies.. no license work to be done. Unless there are future software updates that block it?

Have fun flux, clean, leave no traces of muck!

[1359944032]

Thanks for your reply, the license I am referring to is to open more features. For example, MSO, FRC, etc.

[TheSteve]

Thanks for your reply, the license I am referring to is to open more features. For example, MSO, FRC, etc.

That information is all contained in the original hacking thread here:
https://www.eevblog.com/forum/testgear/dsox2000-and-3000-series-licence-have-anyone-tried-to-hack-that-scope/

[1359944032]

Thank you very much, forgive me for not looking at it seriously .

[oaba]

To save the pads.
Use low temperature solder (Like chipquick, etc. ) first with load of flux. Prior removal.

[EE-digger]

Fearing the shields, I held off any mods for a long time.  Finally did a 500MHz this morning on my 3000T and couldn't be happier.  605MHz BW (0.707) and 750MHz (50%).  Triggers beautifully everywhere but response hits a brick wall just a hair above 1GHz on this mod.

But regarding shield removal, it was a piece of cake.  I used a Pace medium/large chisel and Chemtronics Soder-Wick of the Rosin type.  Solder wicked away easily and the pads cleaned up like new.

Did the component removal with hot air and, against my better judgement, with no bottom heat.  This does take a bit longer when an attached via ties to a plane but still not bad.  This was with a Quick 861DW hot air station, my new favorite.  Super low air velocity is a plus for this tool.  It made my old Pace ThermoJet look like a toy in comparison to the wattage it delivers to the air.

With irons, the biggest shortcoming in SMT removal is inadequate heat and too small a tip, which cannot transfer the heat adequately.  Especially with the higher melting point of unleaded solder.

TheSteve had a good idea with his "undo" kit of parts.  He used separate bags.  Being somewhat lazy, I used double sided tape stuck to a card.  Parts are laid on the tape by channel and rough arrangement.  I let the parts cool first so they wouldn't get permanently "glued" to the card.

[maomi84]

This is amazing！THANKS!  Steve
I `ll do it myself.
I have more question.
1、About K1 relay,Can I use 10dB relay here , such as  a150-10-12 & a150-10-5(Add  dc-dc ),I can only buy 10dB relays.
2、Is calibration necessary after mod.

[TheSteve]

I doubt you can sub in a different value for the attenuator relays and have the scope handle range changes properly.
The self calibration seems to take care of the calibration issues. If this is adequate for your use case only you can decide.

[dew]

There seems to be more differences in in 4000 series hardware. I took some pictures of the area with missing Cyclone IV FPGA and it's surrounding on both sides of the board. Firmware would be puzzled when it fails to find that FPGA that suppose to be there on 1 GHz model. Lots of bypass capacitors and other stuff are missing there as well on 200 MHz model.

[analogRF]

There seems to be more differences in in 4000 series hardware. I took some pictures of the area with missing Cyclone IV FPGA and it's surrounding on both sides of the board. Firmware would be puzzled when it fails to find that FPGA that suppose to be there on 1 GHz model. Lots of bypass capacitors and other stuff are missing there as well on 200 MHz model.

I dont have this scope but I think that FPGA is for one of the options (USB decoding maybe?) so if you dont enable that option, I guess it should work 

[TheSteve]

The 4000 series has the extra FPGA installed on the 1 and 1.5 GHz units.

[EE-digger]

** parts have been sold **

I decided to sell the parts I had collected for a mod of my 3014T.  All are new, from distributors, except for one of the semi's may have been from ebay.

[maxwell3e10]

This is amazing level of hacking! I am wondering if it would be slightly simpler if one starts with a 500MHz 3054T? Would that substantially reduce the number of components that need to be replaced to get to 1 GHz?

[EE-digger]

TheSteve will probably chime in with better info but I don't think there is an easy method of identifying the parts that need to be swapped.  I believe other efforts started with the 100MHz model.

Also, keep in mind that 500MHz units I've looked at actually had a bandwidth around 600 to 625 MHz.  I believe I've seen that the 1GHz scope has more of a brick wall above 1GHz (whereas the lower BW scopes keep going and going, albeit with loss of amplitude.)  Just an FYI.

[EEVblog]

This is amazing level of hacking!

Agreed. I'm impressed that people are willing to hardware mod such an expensive high end scope.

[2N3055]

This is amazing level of hacking! I am wondering if it would be slightly simpler if one starts with a 500MHz 3054T? Would that substantially reduce the number of components that need to be replaced to get to 1 GHz?

Did you read this topic?
Literally in first post by OP there is a link to 500 MHz conversion by Memset where it is explained.
https://www.eevblog.com/forum/testgear/keysight-3000t-scope-1-ghz-hardware-upgrade-mod/msg1245159/#msg1245159
500 MHz conversion is substantially easier.

Sorry my memory serves me badly... That particular conversion was 3000A, not 3000T to 500MHz.

[TheSteve]

The 500 MHz mod for the T series should be roughly the same as the A series however I don't think anyone has ever actually done it to verify 100% the exact components that need to be changed. And yes, it would be much simpler compared to the 1 GHz mod.

[tv84]

Agreed. I'm impressed that people are willing to hardware mod such an expensive high end scope.

After Steve lead the way, for others should be child's play (except me, of course... ).