Agilent 600MHz 54831B hack for 1GHz 54832B possible? YES!

[Jwalling]

So I've just successfully repaired an Agilent 54832B.
https://www.eevblog.com/forum/repair/agilent-54832b-ch1-vertical-errors-(flatness)/

(would love to get some opinions from you folks on a fair refund amount from the seller)

And since I have some 54831B ACQ parts board, I thought I'd take a look around for strapping resistors, since they seem to share the same base board.

A very likely area is shown in the attached pictures. I'm going to attempt to upgrade the 600MHz 54831B on my bench to the 54832B 1GHz model when I get some time (hopefully in the next few days)

EDIT: changed subject line.
Jay

[Jwalling]

The answer is yes!
After removing a single 1K resistor, the 54831B now identifies itself as a 1GHz 54832B.

Jay

[Jwalling]

Vertical amplitude is well within specification. On this particular scope, CH1 had the lowest measured amplitude of the 4 channels. I used a Rhode & Schwarz SMT03 set to 1GHz @600mV amplitude. The second picture highlights the 1K resistor that must be removed.

[TheSteve]

Awesome! Got to love a nice simple hack that gives free bandwidth.

[Jwalling]

An over-all picture of the ACQ board. Also, a close-up of the memory used. Each chip is 2MB in size and there's 8 on each side, so 32Mpts is possible. This one only has option 080 which I believe is 16Mpts. Hmmm...  ?

Jay

[Jwalling]

Awesome! Got to love a nice simple hack that gives free bandwidth.

Thanks. It is a very simple mod to do. Technically, the 1GHz scope cost Agilent (very) slightly less $ to make. 

Jay

[Jwalling]

Final bandwidth test. This time using a high quality low-loss RF cable.
1.26GHz!

Jay

[Nuno_pt]

Nice one, now it's just a matter of finding one on the cheap side.

[Howardlong]

Awesome!!! I have a 54831D, in equivalent time bizarrely it has a bandwidth in excess of 1GHz, but it's ITRO 800MHz in real time. I'll give this modification a go.

Furthermore, what is the fastest timebase since the mod? The kosher 1GHz editions go down to 200ps/div.

[Jwalling]

Awesome!!! I have a 54831D, in equivalent time bizarrely it has a bandwidth in excess of 1GHz, but it's ITRO 800MHz in real time. I'll give this modification a go.

Furthermore, what is the fastest timebase since the mod? The kosher 1GHz editions go down to 200ps/div.

Good question - I hadn't even thought to check before/after.  This one goes down to 200ps/div. Does the 831B/D only go to 500ps/div.?

EDIT: When you have your scope open, can you tell me what the ACQ memory chips are? Thanks.
Jay

[Howardlong]

Awesome!!! I have a 54831D, in equivalent time bizarrely it has a bandwidth in excess of 1GHz, but it's ITRO 800MHz in real time. I'll give this modification a go.

Furthermore, what is the fastest timebase since the mod? The kosher 1GHz editions go down to 200ps/div.

Good question - I hadn't even thought to check before/after.  This one goes down to 200ps/div. Does the 831B/D only go to 500ps/div.?

EDIT: When you have your scope open, can you tell me what the ACQ memory chips are? Thanks.
Jay

It might be a day or two before I get time, I have the wife's family staying with us at the moment.

[Jwalling]

It might be a day or two before I get time, I have the wife's family staying with us at the moment.

Hey now, what's more important!?
We could have a bandwidth contest. Pick the channel with the best performance, and in 1MHz increments (no sense in splitting hairs finer than that), post your results here. 

EDIT: screwed up quoting. Doh.
Jay

[Howardlong]

Family can wait ;-)

Needless to say, scope is now a 54832D, rise time is about 320ps on real time sin(x)/x, and a smidge under 300ps on equivalent time.

Bandwidth is 1230MHz on real time, 1310MHz equivalent time.

[Jwalling]

Family can wait ;-)

Needless to say, scope is now a 54832D, rise time is about 320ps on real time sin(x)/x, and a smidge under 300ps on equivalent time.

Bandwidth is 1230MHz on real time, 1310MHz equivalent time.

Nice. 

It looks like we both just posted at about the same time. What are the odds?
So, your board has 128MB of ram. What is the memory option (if any) on yours?

Edit: And did the 831B/D only go to 500ps/div. before the mod?

Jay

[Jwalling]

Family can wait ;-)

Needless to say, scope is now a 54832D, rise time is about 320ps on real time sin(x)/x, and a smidge under 300ps on equivalent time.

Bandwidth is 1230MHz on real time, 1310MHz equivalent time.

Nice. 

It looks like we both just posted at about the same time. What are the odds?
So, your board has 128MB of ram. What is the memory option (if any) on yours?

Edit: And did the 831B/D only go to 500ps/div. before the mod?

Jay

Never mind on the timebase question, just looked it up in the manual. 500ps/div.
Note to self, stop being lazy...
Jay

[Howardlong]

Family can wait ;-)

Needless to say, scope is now a 54832D, rise time is about 320ps on real time sin(x)/x, and a smidge under 300ps on equivalent time.

Bandwidth is 1230MHz on real time, 1310MHz equivalent time.

Nice. 

It looks like we both just posted at about the same time. What are the odds?
So, your board has 128MB of ram. What is the memory option (if any) on yours?

Edit: And did the 831B/D only go to 500ps/div. before the mod?

Jay

The option I have is MEMORY_64M.

Yes, 500ps is the fastest timebase as a 54831.

Is your trigger jittery at fast rise times in realtime? And for fast sine waves, is the trigger point offset?

[Jwalling]

The option I have is MEMORY_64M.

Yes, 500ps is the fastest timebase as a 54831.

Is your trigger jittery at fast rise times in realtime? And for fast sine waves, is the trigger point offset?

At 1GHz with a 600mV RMS sine wave, the scope will trigger fine without jitter from about +546mV to -434mV. On the second question, I didn't see any problem like that, unless you mean what I just wrote. Maybe running the calibration routine will correct?
The legitimate 54832B behaves the same way, though.

There's an interesting thread about triggering on a sine wave on the Keysight forum about this: https://community.keysight.com/thread/23229

Regarding the PM question about how I figured it out, I had a legitimate 54832B that I documented a repair thread about. Comparing the two ACQ boards side by side made it easy. 

Jay

[Mrt12]

Awesome! it worked, I have a 54832B now :-)

[Noody]

Can you see it? 

http://www.ebay.com/itm/Agilent-HP-Infinium-54832B-1Ghz-4GS-s-4ch-Digital-Oscilloscope-/281890103837

[tonykara]

I just stumbled onto this in 2019, and I'm so glad because it worked on my 54831! 
Jay, I initially looked at your two photos here, and it looks like you took one resistor off in the dual row area, but also installed one in the single row below it. I tried that initially and windows crashed sometime after launching. Then I took off the resistor I added, (so the overall mod was just removing the 1 resistor in the dual row area), and it now comes up as a 54832! So cool!!!! Thanks. Now I need to figure out the changes needed to open up the memory. (I think that sorta documented here somewhere)

[Jwalling]

I just stumbled onto this in 2019, and I'm so glad because it worked on my 54831! 
Jay, I initially looked at your two photos here, and it looks like you took one resistor off in the dual row area, but also installed one in the single row below it. I tried that initially and windows crashed sometime after launching. Then I took off the resistor I added, (so the overall mod was just removing the 1 resistor in the dual row area), and it now comes up as a 54832! So cool!!!! Thanks. Now I need to figure out the changes needed to open up the memory. (I think that sorta documented here somewhere)

You mean the one that looks like a resistor was there, and now is not? Not sure what I was doing there, I can't remember.
But yes, just the removal of the single resistor is needed.

EDIT: And Howard's config is slightly different in that area. I didn't notice that back then...

[analogRF]

Did anybody figure how to enable higher memory options?
Low serial numbers (like mine) do have up to 16Mpts already on board but mine does not have any memory option and only has 4Mpts enabled.
A software key enables 04 and 08 memory options which I really need . Has anybody figured how to do that?

higher serial numbers have the highest memory already installed and need software key to enable them

[bostonman]

I saw the oscilloscope "now identifies itself as 1GHz BW", however, I didn't see where to find this. I'm considering removing the resistor on mine.

From what I can tell, when the scope identifies itself as a 54832, that indicates it's a 1GHz BW due to the model number change, is this correct?

Mine is a 54831M (although, in the 'about' menu, it shows as 54831B), so I assume after I remove the resistor, it will show as 54832B or M?

Reading this thread, it seems this is a true 1GHz BW scope since tests have been done. Do any risks to over driving circuits exist by doing this conversion?

A software key enables 04 and 08 memory options which I really need . Has anybody figured how to do that?

I had read about this upon buying the scope, and I think an option exists in one of the menus to enter some key. I'll have to look into this.

[luckystew]

I have a problem because after performing the upgrade, during calibration i get all channels fail on both vertical and trigger.
Hope that someone can help me with that.

[bostonman]

Sure you removed the correct resistor?