Oscilloscope YOKOGAWA LD1640

[GGMM]

Good morning,
I recovered a YOGAWA LD1640 oscilloscope. It looks very nice on its design and has 4* 200MHZ channels and a 6.4 inch tft screen.
I just got it, it looks a bit worn, but I’ll clean it.
I’m quite capable of completely demonizing it and exposing it.
Of course, there is a problem  because it would be too simple.
It presents a calibration error when starting and also if we try the procedure. (NB: It automatically calibrates at regular intervals)
You have a lot of tests
Memory test gives this error; ACQ_12 Med WA:0 RA:0 WD:1 RD:0
I read on forums that calibration problems on other oscilloscope type tektronic that it could come from the capacitors leakage  on the acquisition board on the 12v  or  5v ?
Your opinion
If you had this problem, give me a lead. Anyway, we’ll have to show for inspection.
cdt

[GGMM]

I add:
The lithium batterry is good. 3.6V

[GGMM]

I found this post very interesting
https://www.eevblog.com/forum/testgear/yokogawa-dl1740-oscilloscope-repair-teardown/

i think there is here some interresting ideas to read.

Maybe there’s a channel or something. I have to dig.

I’ll read all this, but I have to translate
cdt

[shakalnokturn]

Is that the bashed-up one that was selling on Leboncoin not so long ago?

If so, I hesitated on buying, it was just too expensive considering the state of the case.

I'd assume your error message concerns channels 1 and 2, I seem to remember reading somewhere that the RAM multiplexers failed on some Yokogawa DSO's.
Of course the usual advice applies: Check power (then check again at the IC's), find another oscilloscope, look for reset, clocks, stuck buses...

[GGMM]

Hello,
I need to dig into the testing.
With the test we can test each channel one at a time. (I think)
But strangely it is in systematic error for the 4. Amazing because in the posts it is often 1 or 2 HS channels.
It is also strange that the rear outlet tests indicate "not connected".
I need to see if the exit card is present and well connected.
Maybe all this is coming from an HS voltage? Maybe the self-test should tell?
Frankly, yesterday after the carcass was dismantled, I spent only half an hour on it. I see this afternoon.
NB: On the post, it says that this would be a chronic failure of the channel amps. A power surge at disassembly and/or heating.
he preconise to put zeners 5.1v on the 5v and -5v, but do we simply cable them in// on the power supply or provide a resisistance. I need a schematic.
In case I go see if we can put an internal fan or something. I read that some holes are drilled to evacuate heat at the preamp level.
But before I have to make the mistake

I was also wondering, if we shouldn’t do a general reset of the memory stored with the battery.
Do a short circuit of the battery with a small resistance a few seconds (?)  I have not seen for the moment of general reset in the menu
If you have a solution do do that

cdt

[GGMM]

Hello,

I found in the documentation a way to reinit all parameters

All settings will be initialized. ---->>>Power up with the RESET key depressed

to try ....
cdt

[GGMM]

Hello,

This day I completely dismantled the oscillo.
I did a practically microscopic inspection of the components, there is nothing abnormal and no smell of burning. No component "burned"
As the error is on channels 1-2, tomorrow I will reverse modules 1-2 and 3-4 to see if the error is reversed.
But since everything seems normal at first glance, I could investigate a little more.
I will spare myself if it is not necessary to heat a little the CI's  that do not have apparent connections, a little is done on the PS console, because no trace of destruction or deformation. I also tested all capacitors no problem

cdt

[nctnico]

If the self test throws an acquisition memory error, then it is of very little use to swap the channels. The problem is in the acquisition memory. If this is on a seperate board then cleaning the connectors with alcohol and re-assembling the oscilloscope may do the trick.

BTW: don't run the analog front-ends without the heatsinks attached. These can overheat very easely and then you have more trouble instead of less.

[GGMM]

Hello Nctnico,

I will still try because the error indicates a scan bp on channels 1-2 (1module PCB for 2 channels)
If the error is reversed, it will be a pb on this PCB module.  To be seen.
As for the thermal protection, I added the thermal pate on the IC’s and the conductive bearings. This can only be beneficial. After as indicated on another post one can imagine improving the ventilation. Besides the rear fan is too noisy for my taste, in case of success I will look at a silent model.

I see you helped out a 1740, but the components looked pretty damaged.
Do you have the schematic?
You also said that we need 5.1v zeners on +5v and -5v power supplies.
I don’t know if the main power supply provided these voltages, or if small regulators are on the channel board.
If you have the information, I’m a taker.
and to finish how connect the diodes? Simply in // on the  + - 5.1v??

cdt

[GGMM]

hello,
i  got it all back together thinking I could reverse 1-2 and 3-4, but in fact the card is not scalable on this model.
I did a visual check, but I didn’t see that.
I’ve tried to heat the circuit in question, but I’m at the same point.
If the fault is in a component, difficult to see. Also impossible to find the diagram.

cdt

[TomKatt]

Memory test gives this error; ACQ_12 Med WA:0 RA:0 WD:1 RD:0

Memory Test
This test checks the internal ROM. The ROM is functioning correctly, if Pass is
displayed. If Failed is displayed, contact your nearest YOKOGAWA dealer as listed on
the back cover of this manual.

No experience with this equipment, but if I had to guess based on the symbols I'd interpret this as
ACQ_12 - Channels 1+2
Med ?
WA: Write Address
RA: Read Address
WD: Write Data
RD: Read Data

Now, obviously you can't write to ROM so either the memory test checks more than just the ROM or I'm completely off base.  If everything else seems to work correctly and you don't see any kind of pcb issues, it's unfortunately quite possible a memory IC has failed.

[GGMM]

Hello,


I did a visual review of the PCB's again.
nothing special
The error in choosing the memory test is:
Memory test gives this error; ACQ_12 Med WA:0 RA:0 WD:1 RD:0

I think that for channels 1 and 2 this must correspond to the RAMs D45128163G5 (128M Dram 4 bank LVTTL). (if we still follow the logic of the implementation).
For the converter if we always follow the logic of the implantation of the components it must correspond to the one in red.
My question is how to determine the failed component? A ram or a circuit that handles all this.
It’s impossible because the cpu is on top.

I read somewhere that you can fool the converter with a triangular signal. Basically if we can display the triangle but if there are hatches, it would prove that it works, but ram problem. Not easy without a diagram.
Your opinion.

NB: I think Tomkatt was talking about RAM and not ROM

Also one idea,  Do you think there may be an anomaly in the NVRam powered by the lithium battery. 
I was wondering, if it would be useful to cut the + wire of the battery and reset the card by putting a ground wire to empty the capacitors. (?)
Then get back together.
Unless there is a trick (??) with NVRam

 

[GGMM]

I wanted it to be impossible to look at the signals on the oscilloscope, because the CPU board is plugged on top!

[GGMM]

Hello,
I thought I’d find a defectous regulator, but it’s actually good. it would be too easy!! Lol
I repeated memory tests, this gives at least 4 different results. I can not see anything on this card because covered by the cpu.
So, I wonder if a failed memory chipset can give this result (random or not??).
This is where we regret that it is not on supports.
the question of how to identify the component. I can’t rely on the heat, the time I show...

cdt

[GGMM]

Hello,

This afternoon I trained myself to desolder circuits of the same type (54 legs). I took an old memory stick and this is the first time I tried with a hot air gun.
I put 450°C. At the stiffness no matter if the component is still good after. It worked relatively well, but in the environment of the circuit there must be no componsant smd, because it will leave too.
Now, the question is, do we need to change all four or one at a time?
Knowing that already in the lottery I am not too lucky!! LOL.
Too bad no one has the knowledge of this material to guide me.
NB: The chipsets are located.

I hope I’m not wrong, the CI's to change for me are surrounded in red in a previous answer.
Your opinion please.

cdt

[max.wwwang]

Hello,

This afternoon I trained myself to desolder circuits of the same type (54 legs). I took an old memory stick and this is the first time I tried with a hot air gun.
I put 450°C. At the stiffness no matter if the component is still good after. It worked relatively well, but in the environment of the circuit there must be no componsant smd, because it will leave too.
Now, the question is, do we need to change all four or one at a time?
Knowing that already in the lottery I am not too lucky!! LOL.
Too bad no one has the knowledge of this material to guide me.
NB: The chipsets are located.

I hope I’m not wrong, the CI's to change for me are surrounded in red in a previous answer.
Your opinion please.

cdt

The quality of the photos has room for improvement, for gurus (not me) to be able to comment on. 

[GGMM]

Hello,

Yes, the quality of the desoldering pictures is not perfect. The goal was to prove that i can desolder components memory similar to 54 legs.
I watched this video before testing.
 

I managed on an old memory of PC to desolden, so it should be okay.
I made very precise pictures of the environment around each memory in case a component would unsolden.
In addition I ordered the same type of memory and manufacturer and also kapton to protect around.
After the question is: Change all 4 components at once or change one at a time and have a little luck to fall on the first one down.
cdt

[jonpaul]

My experience: The old Yokogawa  scopes 1970s..1990s  are poorly documented (schem in Japan and not public) and any digital Yokogawa scope  has many hard to get IC and proprietary FW.

My take: Reading the OP notes I   doubt if this OP can fix the scope. Even a very skilled tech or eng ma be in similar situation.

My suggestion: Old analog scopes are cheap and easy to find locally (eg Hameg 103 EU5 Hameg 204 EU 25)

The Asian  digital and UB scopes  are low cost,  suggest to drop the repair  project and find a new scope.

Sorry,

Jon

[GGMM]

Hi Jonpaul,

I don’t know what a PO rating is (?PO  = post??)
May  you explain.
Yes, it’s very hard to find technical documentation, but I found everything that is user manual, and also DL1740 technical documentation. No electronic schematic. This can help a little.
 I have also  oscillos  analog PM3055 and a Hameg 312.
Otherwise, I’ll have all the hardware to change the 4 acquisition memories for channels 1 and 2. We’ll just have to test the experiment and see.
If you have more information on diagnostic, you may help.

Ideally, a technical who has worked at the Yokogawa after-sales service should give his resolution diagnosis.
The Korean after sales service  answered me by mail  to invest the card in korea. LOL

cdt

[Haenk]

After the question is: Change all 4 components at once or change one at a time and have a little luck to fall on the first one down.

I'd go with all 4 at once, assembling/disassembling might break additional stuff, so...

(edit: I still have a 1540 around, I found it great as a non-permanently installed scope, the size and weight of a Mac, easily portable and not taking up a lot of space on the bench. It has long been replaced by a LCD scope, but it is really lovely, when it's army of selenoids starts...)

[pcprogrammer]

I don’t know what a PO rating is (?PO  = post??)

He wrote OP which means Original Poster and therefore you GGMM 

[jonpaul]

Rebonjour GGMM, j'habite à Paris 6eme, peut-être sommes-nous voisins ?

Dans et autour de Paris se trouvent des experts en instruments électroniques vintage, qui ont réparé de nombreux scopes, TEK, HP,

1/ PO .......no it is OP = Original Poster, YOU!

2/ Yokogawa: Since 1992  I had good connections to Yokogawa-san , Japan and USA,  we have  WT110, DL1640,  DL7440, DL1740, TA320, TA720.

Despite  connections  we never  obtained any  doc or schema or repair info on these 1990s units besides the operation manuals.

Even replacement of backup battery is a huge hassle.

3/ Fiddling with SMD CPU/RAM/NVRAM is risky and unlikely to fix it.

4/ SMD chip removal and replacement requires the right materials, tools and practice. Easy to damage the PCB or chips.

5/ Only idea is send a note in very basic English to Yokogawa HQ in Japan, ATT TECHNICAL SUPPORT OBSELETE INSTRUMENT

You might  get a schematic or manual of service (in Japanese!)

Amicalement

Jon

[GGMM]

Hello,

Thank you for your answers.
You are right this oscilloscope is very nice and I find that its design is relatively different. Not too heavy and full of features.
I read that we could improve its ventilation by drilling a few holes on the housing and that the fan is quite noisy.
Thanks for your answers on OP, we learn every day.
For disassembly , it’s actually well done and easy to do.
The problem lies in the fact that the cards are superimposed and embossed on the motherboard, this poses a difficulty to see the signals. Surely at the service they must have extension boards for the tests.
You will see on this video that the battery exchange can be done, but indeed there is desassembly.



Don’t worry about disseminating the memory's , I think I can do that by taking a maximum of precautions.

cdt

[GGMM]

Hello,

I got kaptron this day.
I prepared the ACQ card to desolder the 4 memories and also protected the main connector. See picture.
Having never used this product, I did a test on my old Ram bar to see. So I selected the smallest SMD components and put the kaptron on top. See picture
I did a two-minute test at 230°C with full ventilation, then another at 400°C 2 minutes to see.
I’m really surprised, because nothing has changed (that’s the goal)!
Even the bottom of the PCB is no warm, you can put your finger on it without burning.
Well, I’m going to look at the properties of this product, this may come from aerospace origin?
Rest a question: For you the best temperature, duration, and ventilation level (from 1 to 10(max)).  Knowing the chips won’t be recovered.

cdt

[GGMM]

Hello,

https://www.eevblog.com/forum/repair/tds-754d-diagnostic-test-acquisition-fail/

I found this post interesting.
It may be about the same.
For now I don’t know if there is an option to run the test internally in a loop.
For the memory test, a priori, no choice on the "type" of memory to test. It’s all or nothing.
This oscillo also has a USB port, I don’t know if we can write a test program as in this example. (??)
It doesn’t seem obvious to find the failed chip (4 memories for 2 channels)
In short, changing them will take less time, but not yet received.
Track to study can be, see the state of the welds of the 4 rams.
The outage occurred after a long period of non-use.
False contact, oxidation???
But for that you need a hot air station, and no luck, mine is down!! Components ordered. Lol

cdt

[GGMM]

Hello,
For now I can not troubleshoot my DSO because my station is down (I wait for the components), despite everything I wonder about this post and especially towards the end where a script is written to test the memory (?)
see here:

https://www.eevblog.com/forum/repair/tds-754d-diagnostic-test-acquisition-fail/

I really wonder how it works and more, the result of the defectful chip is displayed (Very strong!!)
I don’t know if this indication comes from the app itself from the DSO.
So I wonder if we can do the same thing on the LD1640(?)
My LD1640 has mainly a serial port.
I found on the Yokogawa site 2 programs to emulate the serial port.
See Doc.
Otherwise I have no additional information. I want to avoid any additional damage (Ex: Overwrite the internal program).
It sounds a little complicated, but just in case

cdt

[GGMM]

Hello,
I started changing at least 2 of 4 ELPIDA memories 45128163G5 A75 -9JF
We find the datasheet, but I want the internal schematic if possible. Indeed by comparing to the multimeter the impedance seems different on some outputs DQ. Ex the DQ2 or DQ3 compared to the DQ1/ mass.
 The memory acquisition part of channels 3 and 4 that work have the same impedance  on DQ1 to DQ15.  Maybe there are open collectors or other?
thank's for your help.
I continu my research..I will also look at VCC and VSS because there are several on these chips

cdt

[GGMM]

Hello,

I started changing at least two of the four memories.
And no luck, the result is disappointing.
I may have done something wrong, but actually give an indication.
So I still have an error on the acquisition memory for channels 1-2, but now an error on the SNAP memory.
This would prove that I may be on the wrong track and that this memory would only be used for catching pictures. (??)
The acquisition memory would therefore be elsewhere (Yokogawa specific large flat case)??

The replaced memories have the same references and I checked the welds under a microscope and multimeter )
Components coming from far away... A defectous, fake copy, one can imagine everything

cdt

[coromonadalix]

and you started another thread

[GGMM]

Hello,
Yes I have created another topic. Yes, I understand that this can be complicated to follow
Here is the link:
https://www.eevblog.com/forum/repair/measurement-on-memory-with-multimeter/
but was specific on testing memories...

So:
I come back to this subject because it concerns the DSO Yokogawa DL1640 directly.
As a follow-up to all my tests on memories, I will explain what I did this afternoon.
I changed the U911 memory and did a power-up test.
And there, miracle, the error in 753 in red on the whole screen no longer appears.
All 4 channels appear functional!!!
If I go to the calibration menu, it works perfectly.
---------
But, inevitably, there is something very strange!!
The front panel indicates DL1640, therefore designed for a memory length of 8MB
The config indicates a memory length of 32MB and a model 701620  (See picture) The label on the cover indicates a model 701610 (8mb) in phase with the inscription of the front face (LD1640).
And what’s amazing is that the acquisition error when doing the memory test is on all 4 channels now!!!
The version looks old 1.2       Most recent 1.53
see here:
https://tmi.yokogawa.com/fr/library/documents-downloads/firmware/dl1620dl1640dl1640l-digital-oscilloscope/

The firmware looks identical for both models.
Now I don’t know what may have been "cobbled together"

There is a card reader, so if I want to make an update I have to buy a card. I have to look into how it works?
I’m thinking about what to do next
I’ll get a can practice with (Lots of buttons and functions) LOL
cdt

[shakalnokturn]

So, does that mean that it is more - or less working?

Just curious does your acquisition board have B8050MD or B8050ME model checked?

It could be as simple as a setting in a service menu. How to get to it though... Some older model service manuals could give hints.

Now I don’t know what may have been "cobbled together"

If it looks anything like attached photo it's more "cobbled apart" 

[GGMM]

Hello Shakalnokreturn,

I will show again (Very simple) and note the PCB’s references.
A priori I had not seen a rider to pass 8MO 32 MO
It is true that for my first tests I had put "Zero" the backup battery , but I think it does not play.
There may be a hidden menu. (Key combinations at startup). We find the service manual of 1740, I will look

Yeah, that’s the one that was on E..y. As I had seen on forums, that a calibration error was more due to leaky capacitors, I tried
I restored and painted it in blue except the front (which remains quite yellow) because there are screen prints. LOL.
A priori for my first tests I notice that the screen is divided into 4 zones, one per channel.
It’s relatively small per unit.
I’m wondering if there’s a way to use the full screen for two views and see if we can superimpose two signals besides the Zoom. This is done very easily in analog.
Good to study all functions..

[GGMM]

Hi,

It seemed to me that I had noted this information.
CPU  B8050MA
ADC  B8050MD

cdt

[GGMM]

Hello,
I’m coming back to you on this DL1640.
It seems to be working, but I find that there is noise on the displayed signal.  See the video.
Since my repair all channels are ok but before only channels 3 and 4 were functional ..... but I didn’t pay attention if there was noise!
I don’t have experience with DSO, but you can confirm that this is not too normal.
I have at this moment only one hypothese of work: The operation seems to have become operational at the change of the 3rd memory of a bank of 4.
The references are the same, but I’m sure they’re not the same lot fabrication . I don’t know if that’s going to have an impact or if the fourth should be changed for that
Unless: Bad weld or something.
Your advice is valuable.

cdt

[shakalnokturn]

Although your photo isn't saying much as we don't know the vertical sensitivity that looks normal to me for a DSO.
Have you re-attached the shielding on the frontend / ADC section?

Try coupling all the inputs to GND and turning vertical sensitivity right up that should give you a better idea on abnormal noise.

For the acq. RAM if the replacements are at least as fast as the originals there shouldn't be a problem, if there is a problem I'd expect it to show at highest sample rates if anywhere.

If you bought a set of 5 RAM IC's on eBay or Aliexpress you have no idea what you really bought, just keep that in mind if it no longer works sooner or later.

[GGMM]

Shakalnokturn,

Thank you for your answer.
For purchases on A...x or E...y, there are things I’ve had for ages that work.
The advantage is that there are no minimum purchases like on other sites that sell components. Well, in short...

To talk about something else, YES it made me tilt after the post, the DSO is open, so the shielding under the ADC is not set for now that can play.
I will see this

I put you the PCB references as requested. (See above)
 If we can lift the doubt for the RAM 8 or 32 M/ channel. Please give me your answer
Otherwise, I did not find a soft option of choice between the 8 and the 32.
There is an option by holding the reset button and do the power up, but this is more to reset the internal user settings.
So it remains a mystery between this model and its size of memory in the acquisition card, but there is always an explanation.

NB: in a menu you can choose an acquisition sockage option that goes from 10K, 100K, 1M, etc and 32M (single), so it seems consistent.
Ha! ,If technicians from YOKOGAWA could read our posts...

Cdt

[GGMM]

Hello,

Regarding noise, watch this video.
How the passage of 40 years of analog makes you ask questions about digital (especially noise). LOL



cdt

[GGMM]

Hello

Victory, everything goes back to normal
The memory test is good
The overview system indicates a 7610 model
By changing the 3rd memory, there was a short micro circuit between pins 42 43 of the chip

The subject can be closed.
cdt