HP 54601A 100MHz Oscilloscope Teardown

[balu]

Hi Eevblog’er,

I have made a tiny teardown of my HP 54601A 100MHz oscilloscope to find out the production date of the Delas DS1220A non-volatile SRAM. And yes, I have a battery lifetime issue and must replace the SRAM (or battery) in the near future. It is an absolutely easy to use gadget and I will use this scope in future too. 

[balu]

And some photos of the screen with some menu functions displayed.

[mskobier]

Balu,
     I recently came into possession of the same model oscilloscope. I agree, one of the simplest scopes I have ever operated. Push the auto button turn on the cursers and the scope does the rest. According to the boot screen, it is a 1992 model. Does this oscilloscope suffer from the same problem that the Tektronix 24xxB series does if the battery in the Dallas NVRAM dies, you loose all of the calibration settings and must go through a lengthy recalibration to be able to use it again? Mine currently works fine, but I am already in the process of replacing the NVRAM in the Tektronix 2445B, so should be able to remove/read/program and reinstall this chip fairly easily.

Thanks
Mitch

[Firebird00]

Hi.

I know this post is too old, but I need some help with this scope, so maybe you can help ?

I have the same unit 54601a for some time, and one day I get the "Checksum error" message on the screen. After that day the scope did not worked anymore.

I've changed the NVRAM and put a socket to make a easy job in the future. The scope now don't show errors but the channels (all 4) stopped working. When I put a signal in any of the four channels I get always the same message "No signal found".
I see the signal in the screen but can not be triggered, the signal don't stop, Tried all the possibillities with the trigger menu, nothing changes.
I think that this scope lost the calibration. I tried to do a Default calibration procedure many times but always I get a Hysteresys Error in one of the calibration steps and the calibration procedure ends without complete the process.
I saw in this post that you wanted to make a backup of the NVRAM to prevent the same problem I have, so I want to ask you if it's posible you send me a copy of that file, that allows me to burn another (or the same) DS1220 and with luck this scope will work again.
I'm asking that because is too difficult to find people with that scope, and more difficult people that wants to share information like a Dump of the NVRAM.
Thanks in advance.

[Deckert]

Did you try to load calibration defaults?

From the HP Service manual:

1. Check the rear panel DC CALIBRATOR output level.
If you are not sure how to check the DC CALIBRATOR, see "To check the output of the DC CALIBRATOR," on page 4–6.

2. Load the default calibration factors.
a. Set the rear-panel CALIBRATION switch to UNPROTECTED (up position).
b. PressPrint/Utility , then press the Self Cal Menu softkey.
c. Press the Load Defaults softkey.

Manual available here:
http://www.cube.co.za/~tva/workbench/hp54602b/HP-54602B-Oscilloscope.pdf

It covers the series.

--deckert

[SingedFingers]

Also schematics here for that unit: https://archive.org/details/hp_54601A_Schematics

Been after one of these for a while as they're a step up from an analogue but not too far into digital territory.

[Firebird00]

Thanks for your replies.

Yes, I do the Self calibration a lot of times, but always I get a Failed error in the Trigger Hysteresis point, and aborts the calibration process, so the scope still uncalibrated. For the calibration I'm using an RG58 (50 Ohms) cable with two BNC, connecting one end to the back (calibration point) of the scope and the other end to the channel that the scope asks, first the 3, 1, 4 and the #2.

Also I've downloaded the SM and schematics, and traced the signal applied until it arrives to Analog MUX 1 and also THold 1 IC's (on the block diagram), after that I don't know what to see, too many digital things for my knowledge.

[balu]

Hello Firebird00,

I will upload (needs one or two days to be avilable) the NVRAM images as .bin and .hex files to this web site:

http://www.ko4bb.com/getsimple/index.php?id=manuals

I think you will find the ZIP file with the NVRAM image files at:

http://www.ko4bb.com/getsimple/index.php?id=manuals&dir=HP_Agilent/HP_54601_Digital_Storage_Oscilloscope

I replaced the old NVRAM and make a copy of the content. I have hat no problem after the replacement with a new NVRAM.

I hope the copy of the NVRAM will help

[Firebird00]

Thank a lot Balu.

I'll wait for that images,... Not sure if that will solve the problem of my scope, but I lost nothing trying, who knows?
Also, I already have the same schematics you uploaded, from 54600 2 channels scopes. Basically are the same scopes, and at least we have more than nothing.
Did you do the Self Calibration after the nvram swap? I suspect that when I do the self calibration and fails in Hysteresis point,  it's because no signal is detected by the Trigger Hold IC, but I'm in doubt if I'm using a correct coax cable for that purpose. I'm using the RG58C/U that is not too bad, but I'm not sure if a not-too-bad coax cable can produce fails like that...
Another question is what nvram did you put, the new one? I put the DS1220AB-120. I'm sure this one is so far faster tan the original one, but my doubt is with the numbers at bottom of the IC, it shows 0031D 097434. I suspect that the 0031D is the manufacture date, Year 2000 week 31, is it? If that is correct, then I buyed an 16 years old IC ? If so, then I have to buy a new and fresh nvram.
Be continued...

[balu]

Hello Firebird00,

There was no need to use the self calibrtaion after replacing the NVRAM in my scope. Because I just copy
the data from the old one into the new NVRAM. I can't find a new DS1220 in Germany and use the M48Z02-70PC1
from ST. This NVRAM works fine in my scope

M48Z02 see: http://www.st.com/content/st_com/en/products/memories/nvram/zeropower/m48z02.html

[balu]

Hello Firebird00,

Attached to this replay you will find the NVRAM image file as .bin and .hex file as part of the ZIP

I use the TL866CS Universal Programmer to read the data from the old NVRAM and program the new one.

I hope after updated your NVRAM the scope will work
 

[Firebird00]

Hello Firebird00,

There was no need to use the self calibrtaion after replacing the NVRAM in my scope. Because I just copy
the data from the old one into the new NVRAM. I can't find a new DS1220 in Germany and use the M48Z02-70PC1
from ST. This NVRAM works fine in my scope

M48Z02 see: http://www.st.com/content/st_com/en/products/memories/nvram/zeropower/m48z02.html

Ok Balu. That's sound logical, you didn't loss your calibration data like me.

I've found the original DS1220AB-XXX in Mouser
http://eu.mouser.com/Search/Refine.aspx?Keyword=ds1220
So I think I would preffer to buy the same reference as the original. I'll do that and tell what happens

[Firebird00]

Hello Firebird00,

Attached to this replay you will find the NVRAM image file as .bin and .hex file as part of the ZIP

I use the TL866CS Universal Programmer to read the data from the old NVRAM and program the new one.

I hope after updated your NVRAM the scope will work
 

Thank you very much Balu.
I saw in Amazon and EBay some of these programmers. I think that the most simple (without additional sockets) will serve for my purpose, because the DS1220 will fit directly to the programmer, without need for aditional or converting sockets, is that correct?

[balu]

Hello Firebird00,

You can use the M48Z02-70PC1 as a replacement for the DS1220. The M48Z02 is cheaper and based on the datasheet from ST: "The M48Z02/12 ZEROPOWER® RAM is a 2 K x 8 non-volatile static RAM which is pin and function compatible with the DS1220.".

I have hat no problem to read the old DS1220 and program the new M48Z02 with TL866CS Universal Programmer (no need for a adapter)

[Firebird00]

Hello Firebird00,

You can use the M48Z02-70PC1 as a replacement for the DS1220. The M48Z02 is cheaper and based on the datasheet from ST: "The M48Z02/12 ZEROPOWER® RAM is a 2 K x 8 non-volatile static RAM which is pin and function compatible with the DS1220.".

I have hat no problem to read the old DS1220 and program the new M48Z02 with TL866CS Universal Programmer (no need for a adapter)

Hi Balu.
Bought today in Amazon the same programmer as yours, TL866CS, It will arrive between March 23 and April 1. Price is 56,99 Euro and come with 4 adapters. The price of the same without adapters is 51,80 + 4,87 for shipping, so for 32 cents of euro more I have 4 adapters, not bad uh?
I will try to write the nvram I have here, if this don't work or have any issue then I will buy a new one
At this point I have to wait until the programmer arrives, and continue working on this scope
I will post the news.
Thanks a lot 

[Firebird00]

Hi.

I've received the programmer today, burn the file you sent (both HEX and BIN) to the nvram and the problem still there, no trigger at all.
Tried again the calibration process and it fails in the Trigger Histeresys point. I can see the signal apllied on any channel on the screen, but no trigger, also pressing Autoset I get always the same message "No signal found" on all four channels.
Seems that the signal applied is lost in any point of the board, before the trigger, but arrives to other parts that can be processed and showed in the screen.
I've removed and put back the socket of the nvram again, to discard any issue with bad solder point.
All the functions and buttons works (vertical, horizontal, trigger modes, etc.)
All setting remains there when power off and on again, so seems the new (old) nvram I have is not bad.
Any help will be welcome.
Thanks Balu for all your effort helping me with that problem.

[JFJ]

I can see the signal apllied on any channel on the screen, but no trigger, also pressing Autoset I get always the same message "No signal found" on all four channels.

I had the same problem and error message with a HP 54600B. I traced the fault to a failed TL071 Op Amp that's used to buffer the output from the trigger source multiplexers. The TL071 is marked as U43 in the trigger schematic (section 6 of the 54601A CLIP package).

In the absence of any input signal to the oscilloscope, the voltages on pins 2, 3 and 6* of U43 should all be close to zero, with respect ground. My TL071 had failed with its output (pin 6*) stuck at the positive rail voltage.

If your U43 is good, then check U12.1 (AD96687 Ultrafast Comparator). It compares signals from U43 and a DAC (which sets the trigger level), to produce a stable trigger signal.

(*Corrected pin number).

[Firebird00]

JFJ, I've read these voltages on U43:

Pin 2 -> +4,28V
Pin 3 -> +4,87V
Pin 5 -> -4,84V

No signal applied on channels, and loaded Default Setup before taking measures.
The schematics says the voltage on U9 pin 20 (AD9300 Multiplexer) should be zero, and I read +4,88. If this pin is the Multiplexer output, that goes to pin 3 of U43 trough R150 (10K), maybe the problem comes before U43? Voltage drop on R150 is about 4-5mV (+4,8752 -> +4,8708)
What Do you think ?

Thanks for your help.

[JFJ]

I've read these voltages on U43:

Pin 2 -> +4,28V
Pin 3 -> +4,87V
Pin 5 -> -4,84V

Sorry, I misread the output pin number from the schematic. It should be pin 6, not pin 5:



U43 is configured as unity gain voltage follower - its output (pin 6) is fed directly back to its inverting input (pin 2). So, the output voltage has to be the same as at pin 2 (4.28V). That voltage will drop to about 3.5V at the emitter of the common collector transistor (part of U15) that supplies the inverting input of U12.1.

3.5V is far greater than the voltage that is applied to the non-inverting input of U12.1. Consequently, U12.1 can never switch to produce trigger signals.

... maybe the problem comes before U43? Voltage drop on R150 is about 4-5mV (+4,8752 -> +4,8708)

If the voltage is lower on the U43 side of R150, then source of the problem is likely to be before U43. In which case, a good place to start would be to check the voltage at test point 9:

[Firebird00]

Looking today on the schematics I saw the mistake of pin 5 and 6 of U43. As you said, I have the same voltage on pin 6 and pin 2 of U43.

Now I'm working on it, trying to find the source of the problem.

TP9 is the output of U9 pin 20, and I have there +4,88V, as I wrote yesterday

 

The schematics says the voltage on U9 pin 20 (AD9300 Multiplexer) should be zero, and I read +4,88. If this pin is the Multiplexer output, that goes to pin 3 of U43 trough R150 (10K), maybe the problem comes before U43? Voltage drop on R150 is about 4-5mV (+4,8752 -> +4,8708)

I have zero volts on the inputs of U9, pins 3, 5, 8 and 10, +4,88V on pin 20 (Output) and +7,0V on pin 19 (Bypass)
Yes, the voltage un U43 side of R150 is lower, some millivolts.
Also, next to TP9 is C58 (0.01uF) capacitor. I have +4,85V on both sides of that capacitor. If capacitors blocks DC voltages I don't understand why I have these reads, maybe that capacitor is bad also, or is the main problem?
Thanks

[JFJ]

TP9 is the output of U9 pin 20, and I have there +4,88V, as I wrote yestarday

Oops! Preoccupied with my first mistake, I made another (award self zero points for observation).

... next to TP9 is C58 (0.01uF) capacitor. I have +4,85V on both sides of that capacitor.

The voltage on both sides of C58 will be the same when the trigger is DC coupled. That's because C58 - the AC trigger coupling capacitor - is shorted-out by pins 14 and 15 of U10 (DG211 - Quad Analogue Switch), when DC coupling is selected.

If you change the trigger coupling to AC then the U10 switch should open, and C58 will be the only connection between TP9 and R150. That may allow you to determine which side of C58 is connected to the source of the 4.85V.

[Firebird00]

TP9 is the output of U9 pin 20, and I have there +4,88V, as I wrote yestarday

Oops! Preoccupied with my first mistake, I made another (award self zero points for observation).

... next to TP9 is C58 (0.01uF) capacitor. I have +4,85V on both sides of that capacitor.

The voltage on both sides of C58 will be the same when the trigger is DC coupled. That's because C58 - the AC trigger coupling capacitor - is shorted-out by pins 14 and 15 of U10 (DG211 - Quad Analogue Switch), when DC coupling is selected.

If you change the trigger coupling to AC then the U10 switch should open, and C58 will be the only connection between TP9 and R150. That may allow you to determine which side of C58 is connected to the source of the 4.85V.

Dont worry about the mistakes. You are doing a good job helping people (like me) and that is the most important thing.

I switched the trigger from DC to AC, and now I have +4,88V only on the U9 (pin 20) side of C58, and also I have close to zero volts on U43 (OP amp) pins 2,3 and 6, is that correct?.
That information tells you something good? I'm starting to think the U9 is the source of my problem, but I'm waiting for your veredict
Thank you a lot.

[JFJ]

... I have close to zero volts on U43 (OP amp) pins 2,3 and 6, is that correct?.

Yes. Only an AC signal can pass through C58 to U43. Whatever voltage is applied to the non-inverting input of U43 will be replicated on its output. It couldn't quite achieve that when 4.87V was applied, because it was too close to the +5V supply rail.

I'm starting to think the U9 is the source of my problem...

You have already established that the four inputs to U9 are at zero volts. So, unless U10 is leaking its +5V supply to pin 15, U9 has to be the prime suspect.

One way of determining whether it is U9 or U10 that is faulty, would to remove the interconnecting resistor (R196). Though, you might be able to achieve the same thing, without removing R196, if you can measure a voltage difference on either side of the resistor (the highest side points to the faulty chip).

[Firebird00]

Ok.

I took this measures without removing R196. I don't have experience yet removing with SMD components.
I had to switch the trigger coupling back from AC to DC because the less significants digits of the DMM didn't stop moving, then with trigger in DC mode I have read on R196 on the pin that points to U9 about 2-4mV more than the pin looking at U10, that is +4.9042V to U10 vs +4,9056 to U9
If the faulty chip is U9, then I will call a good friend to do the swap for me...but meanwhile I can do some research...
So I think your veredict will be to replace U9 ?
Thank you.

Update today March 30: I've ordered a new AD9300KP from the USA. With luck I will receive it in less than two weeks. I will replace it and see what happens.
Another question I have, what do you think about the +7,0V on U9 pin 19, multiplexer bypass? Is that correct or is bad also like the 4,8V in the output pin 20?

[JFJ]

... what do you think about the +7,0V on U9 pin 19, multiplexer bypass? Is that correct or is bad also like the 4,8V in the output pin 20?

That appears to be correct. The voltage on pin 19 of my U9 multiplexer is +6.9V.

So I think your veredict will be to replace U9 ?

If you want to be absolutely certain that U9 is faulty - without un-soldering any components - then you might consider cutting pin 20 of its PLCC package. That would, obviously, completely isolate the multiplexer's output, but it could be reversed with a blob of solder. Cutting the PLCC pin high up should eliminate most of the risk of damaging the PCB. However, you would need a very fine and sharp pair of side cutters.

As a bonus, isolating the U9 output would allow you to check the trigger circuitry at U43 and beyond. The 'Line' trigger signal is not multiplexed by U9. So, if all is good from U43 onward, selecting 'Line' as the trigger source, setting the time-base to 5ms/div and a channel to around 1V/div, you should be able to display a stable/triggered sinusoidal waveform (the mains power signal that's picked-up by your body) by simply touching the probe tip with your finger.


BTW, I'm sorry not to have responded sooner, but recent security upgrades to this forum's server prevented me from posting. I still prefer to use an old Windows XP PC for connecting to the Internet. The version of Chrome that still supports XP, isn't compatible with the SSL/TLS protocols that are used to provide secure connections to EEVblog pages. I was about to install a current version Firefox (which does still support XP), when I noticed that the http/https redirect problem had been corrected:
https://www.eevblog.com/forum/news/server-ssl-upgrade/msg1178017/#msg1178017

My old PC lives to fight another day