Quick repair project - Marconi 2024

[chicken]

PS: The low voltage on AM Input leads me to a few additional theories (warning, I don't know anything about control loops or RF gear):

• The RF level is the issue and the MCU can't pull AM Input any lower. But I haven't figured out yet, where the MCU would determine the absolute RF level.
  
• The low AM Input causes the issue. This could due to
  a) a broken DAC, which I find unlikely as AM modulation works. Though there are multiple DAC outputs involved, maybe one just controls the RF level.
  b) a calibration issue, i.e. the "steady state" output of the DAC should be higher.
  

[TK]

@chicken, thank you for the detailed reply.  I changed a bunch of TL074C opams today, same result.  My unit PS measures +11V and -11.37V, so I suspect there is some sort of imbalance creating a lot of false.  I removed a 220uF cap after TIP42 (-11V) and the error goes away, output power is set at -2dBm (it used to be -137dBm)... no RF output activity, but no errors.  I will continue playing with the +11V and -11V supplies.

EDIT: I forgot to reconnect the cable from the CPU board to the RF board, that is why it had no RF output activity.  I am now getting a lot of 501 and 502 FRAC/N errors like the original post, but my control board does not have IC207, it has IC14 that is a MAX274BCWI (Active Filters 4th & 8th Order Continuous-Time ).  I am not sure if my board is newer or older than the one from OP.

2nd Update: the 501 / 502 FRAC/N errors go away when calibration 100 is executed.  Quote from Service Manual:

Pre-steer values are derived by an automatic calibration routine (Cal routine 100). If pre- steer problems are suspected, running this routine will cure the problem if no other hardware problems exist. If the routine cannot complete the calibration due to a hardware malfunction an error message is displayed to help identify the problem.

[TK]

I think I found an error on the Marconi 2024 schematics (Marconi_2023A_2024A_2025_Service_Manual.pdf, page 248, Level Modulator AA1 Module Schematics sheet 6 of 9).  The inputs to the opamp TL074C(a) are inverted.  I checked on the board and what was drawn as going into (-) pin 2 goes to (+) pin 3 and viceversa.

[chicken]

Good catch! I can confirm that in my unit as well as Dave's teardown:
https://www.flickr.com/photos/eevblog/7004690115/in/album-72157629274319878/
(ST chip in the center)

[chicken]

Actually, in the service manual that I have (2023 & 2024, (c) 1998), the schematic is correct.

[TK]

@chicken: My manual is 2023A 2024A and 2025 service manual and it has many errors.  I cannot find the 2023 2024 service manual you have, did you get it online?

Found it

[GB97816]

Hello,
sorry that I put in here. Does anyone of you have a binary file of IC5 of IFR2024 (EPROM 27C4001) and could email me this?
In my device, the block seems to be broken, I see in the hex file destroyed pleadings.

Thanks and Regards,
Georg

[gby]

I just got a used Aeroflex/IFR/Marconi 2025 sig gen from fleabay.  It seemed to work at first, but after 5 days of being off I pushed the power button, the display lit up with blue background, and then no text ever.   After lots of time/power cycling with still no display text I figured I had a repair project....

When I opened it up everything looked as expected and no sign/smell of burnt electronics.  Then I took the plate off the power supply section and got a big surprise.  There was not a single ac in to multiple voltage out supply board like in the similar 2023A EEVBlog tear down, 2024 pictures in this forum posts.  There were 4 individual metal boxed ac in to single dc voltage out supplies crammed into there!   So, someone had done a complete home brew power supply replacement.  See Picture.

I then checked the power supply voltages and got +24V, +12V, -12V as expected but the +5V was only 2.45 V.  Even worse, the current was from the device into the +5V supply (backwards flowing).  So, the 5V was gone.  Since the 5V supply runs the processors it makes sense that there was no text.

Next step is to get a working 5V supply and see if that brings the unit back to life.  The existing substitute supply was labeled 5V, 3.8 A.

My question to the group is how much 5V current does this unit need/use?? 

I could not find any specs in the Service Manual nor written on the original power supply PCB pictures I found from tear downs.

Could someone take a look at the supply board carefully and see if any load current capacities are on the silk screen or if you could hook a current meter like the UT210E around the 3 orange wires and measure the load current while operating?

Thanks.

[PA0PBZ]

My question to the group is how much 5V current does this unit need/use?? 

Nice surprise power supply you have there! Can't you hook up a bench PSU to see if the thing will work and to find out the current consumption?

[Andy Watson]

Table at the bottom of 1-39, in the serevice manual lists the "instrument requirements" of:
+5V  2.8A
+12V  2.3A
-12V  0.6A
+24V 0.6A

[gby]

Wow, I must have skimmed around the service manual for quite a while and could not find the current ratings.  So, thanks Andy for the pointer.  In my service manual pdf the supply chart is on page 1-40.  Who knew the instrument requirements would be in a table documenting the optional DC power version.

PA0PBZ I thought of hooking up a bench supply to continue the debug and finding the consumption.  But, I am worried about power sequencing when turning on.  Some multi-supply systems like the supplies turned on in a certain order and/or within a certain amount of time.  Of course, my unit's home brew supply probably lost all supply order sequencing compared to the original...

I will update this thread when I get time to spend more time on it.

[gby]

It took me a while to get a new supply since I ordered a replacement one from China.  When I checked the new supply outside the instrument I thought I would check the old one I took out.  Well the old one seemed to work.  1 month ago in the instrument it put no voltage out.  So, somehow the old one is flaky.

I put the new one in and the unit powered right up with the display showing numbers as it should.  I was able to set carrier frequency and amplitude.  For reference the 5V load was 1.9 A.  Certainly a weird/non-standard supply for this generator but at least it works again.

Now I have figure out the modulation modes and check those out.

[corvo]

Hi,

This is a simmilar problem but for a different equipment series:

I'm looking for informations about one Display for the equipment:

 MARCONI INSTRUMENTS 10KHz - 1.35GHz signal generator 2030 (The Serial Number is 119329/044 Made in England).

The Display is:

 EPSON ECM-A0310

 (9006C0U)

Assuming it is a discontinued part:

Where can I find the data sheet?

Has anyone found a direct replacement?

Has anyone adapted a diferent display to this design?

Has anyone added new hardware to adapt a new display for this generator?


Running the test procedure it shows up the message:

**Can Not Setup Link Between Processors**

and the display shows some corrupted irregular half columns and lines as well.

I have checked all connections from the processor board to the display controller board and it matches as supposed.

 The controller board has the controller IC SED1330FBA

https://www.lcd-module.de/eng/pdf/zubehoer/sed1330.pdf 

The pinout of the display is:


PLFM
 
1 plus 5 V
 
2 GROUND
 
3 VLCO
 
4 LP
 
5 FR
 
6 NC
 
7 NC
 
8 DIN
 
9 MSXCL
 
10 NC
 
11 D0
 
12 D1
 
13 D2
 
14 D3



I imagine a product: Display Box Interface, to convert obsolete displays to new ones.

If it doesn't exist, maybe with the information for the existing parts, we can build one.

Here is another place of information:

https://www.vintage-radio.net/forum/showthread.php?t=153435

https://groups.yahoo.com/neo/groups/marconi_test_instruments/conversations/topics/6244

Thank you.

[LogicalDave]

If anyone is looking for one of these, I have one for sale that needs some repair on eBay; I just don't have time to pull it apart and repair it.  The ad shows the details of what works/doesn't.
https://www.ebay.com/itm/184423645670

[harha]

Hi
I have the same problem with level 2 password for my 2025--could you possibly explain how you sniffed the password from the eeprom?
Regards
Hardy

[grumpydoc]

Hi
I have the same problem with level 2 password for my 2025--could you possibly explain how you sniffed the password from the eeprom?
Regards
Hardy

Also PM'd

In principle it was easy, trigger a 'scope on the EEPROM read/output enable, go to the level 2 lock screen, type some digits hit enter and start the 'scope capture - read the data bus when the firmware accesses the stored password, convert the 24(or possibly 32)-bit value read from the EEPROM into a decimal number and that's the unlock code.

In practice there are a few things that make it less straightforward than that - but not much. On the 2024 the reads that you want are relatively speaking quite a long time after you hit "enter" on the keypad and there are a couple of random, unrelated reads first - but the ones that you want are fairly obvious when you see them. A 'scope which can count triggers before starting the acquisition would make life easier as would doing it with a logic analyser which can capture the data bus 8 bits at a time (I literally had to do it one bit at a time over several attempts).

But the 2024, at least, just reads the number from memory - no clever encoding or encryption.

[George Edmonds]

Hi Harha

I now have the ability to obtain super user passwords again, please let me have the instrument serial number from the label on the rear and the serial number displayed on the LCD screen at startup by private message please.

George G6HIG Dover UK 

[harha]

Thanks a lot.
Have also got an 2025 with lost level 2 password--problem solved  with help from greg edmons-grumpydog and viavi.
Have calibrated it but found a very strange thing!!!There seems not to be an calibration routine for 9khz-10Mhz(bfo range).
Anu clue to cal this range?
Regards
Hardy

[rogersstuart]

Hi
I have the same problem with level 2 password for my 2025--could you possibly explain how you sniffed the password from the eeprom?
Regards
Hardy

Also PM'd

In principle it was easy, trigger a 'scope on the EEPROM read/output enable, go to the level 2 lock screen, type some digits hit enter and start the 'scope capture - read the data bus when the firmware accesses the stored password, convert the 24(or possibly 32)-bit value read from the EEPROM into a decimal number and that's the unlock code.

In practice there are a few things that make it less straightforward than that - but not much. On the 2024 the reads that you want are relatively speaking quite a long time after you hit "enter" on the keypad and there are a couple of random, unrelated reads first - but the ones that you want are fairly obvious when you see them. A 'scope which can count triggers before starting the acquisition would make life easier as would doing it with a logic analyser which can capture the data bus 8 bits at a time (I literally had to do it one bit at a time over several attempts).

But the 2024, at least, just reads the number from memory - no clever encoding or encryption.

I'm trying to do this right now. I have my logic probe connected to the AT28C256 EEPROM and I'm triggering on OE. There's lot of traffic on the bus. Constant traffic. I'm using a Rigol MSO5074 so I can zoom pretty far out to see the pattern. I'm not seeing much if any change when attempting to get it to accept a code. The pattern changes a lot when selecting different menus. It's possible that this may be a different firmware revision than the one you were using. In any case, it's pretty much undecipherable. I'm looking at a dump right now and I see some patterns and I've tried a few sequences but most of it is gibberish. I'm not even sure if any of it is ASCII or not. If it isn't then there's basically no chance that I can figure this out unless I trigger on the button press somehow and really narrow down the window. Even then, I'd need to do some data analysis. Write a script. That sort of thing.

Am I probing the right chip? If you can think of anything please let me know. I'm going to shoot George a message as well. It feels like giving up but I'm better off getting it done easily if possible.

[grumpydoc]

Sorry for not replying sooner - I only get time for an occasional scan of EEVBlog these days.


The AT28C256 should be the right chip but trigger on OE *and* CE being active - OE is just connected to the CPU RD line and will be toggling all the time even when the EEPROM is not being selected.