EEVblog Electronics Community Forum
Electronics => Repair => Topic started by: Jontm on May 24, 2020, 06:06:13 pm
-
Hi!
I got my hands on a LeCroy LT264M that have spent the last 8-10 years in a semi open shipping container exposed to the elements :o
All the other equipement i found in this container seems to be working ok, but the scope has some problems powering on. Fan turns on and standby led light ignites then i press the power button, but the display does not turn on. If i turn it off and try to turn it on after 5+ minutes, the screen turns white for a second before turning off and again getting stuck on the standby led.
From what i can tell the standby light indicates start of self test. I guess it does not pass the self test and gets stuck.
It did start one time, but i have not been able to reproduce this. When it started, i was able to run a test on all the channels to ensure that they where working ok. Everything was working ok. Except for the fact that the next time i tried to turn it on, it gave the same short white flash on the screen and stuck on standby led.
I have opened the unit up and it looks mostly ok without any obvious signs of damage or black spots. I have measured all the voltage rails and they looks to be ok both on the CPU board and the main acquisition board. Tried reseating all the ribbons and connectors, including the flash card and SD-RAM.
I have a feeling that the issue is on the CPU board, but i have not been able to pinpoint it yet
I have focused my main efforts on this circuit that should control voltage to the display, picture attached.
From what i can see, the service manual for this unit does not contain a schematic for the CPU board. I found the schematic for the CPU board on the waverunner 1, but it seems like the VGA circuitry for power to the display is slightly different on the waverunner 2 series. I have however identified that Q10 (and Q21) does turn on for a fraction of a second when i hit the power button, before turning off again. I have however not been able to identify where the signals to turn on and off Q10 comes from, it does not go to the IC86 as on the waverunner 1 schematic.
Anyone have any experience with this scope or perhaps have any ideas where to look next ? Is there any way to bypass the self test or find out what step in the self test it fails on ?
-
One of the first things to check when experiencing intermittent power-up fails other than PSU voltages and ripple on LeCroy scopes (among other TE) is the RTC backup battery voltage.
-
One of the first things to check when experiencing intermittent power-up fails other than PSU voltages and ripple on LeCroy scopes (among other TE) is the RTC backup battery voltage.
Thanks for the input. I checked the battery voltage and it looks fine, just over 3 V.
Is there any point in trying to connect to the serial port ? Maybe it posts some init logs to serial when i try to boot it ?
-
I have done some more work on this and it seems like there is no signals on at least the external 232 port of the scope during boot. At least with the baudrates that i have tried with.
Looking further into the CPU, it seems like that the scope has a JTAG connector with some perhaps useful signals for further analysis:
NSRESET - Measured to be at 5 V
NHRESET - Measured to be at 5 V
NCHECKSTOP - Measured to be at 5 V
Based on this i would assume that the CPU is at least not in a checkstop condition ? And not in a hardware or software reset condition when the scope is stuck with black screen and the standby light illuminated. This might not be of any use or relevant to check, i am unfortunately not familiar with embedded processors and what might be relevant to check.
Is there any other relevant pins to probe to assess if the processor has started properly and is operational ? Are anything specific known about the power on self test ? Such as steps the processor go though and possibly any probing locations that might be relevant ?
-
Considering the scope's history you may want to inspect for corrosion on the FPC and SDRAM contacts.
In normal conditions the standby LED comes on during splash screen then goes off, blinks briefly during the single beep when the graticule is displayed and stays off after that.
If screensaver is enabled the standby LED comes back on while screen is disabled.
Check that no reset is stuck for some reason IC4 (CPU), IC86 (LCD power, VGA reset), IC2 (Bus control).
That CPU clock is running IC3, IC2.
You should be able to use the unpopulated SW1 if you need to make repeated hardware resets.
After that probe around for various bus activities on reset that may point to where it gets stuck.
The CHIPS VGA controller is known to fail on the LeCroy LC series but those do run hotter than the LT.
-
Thank you for your valuable input and knowledge. The scope was in much better condition visually than what i expected based on the prior history. No visible corrosion on any contacts and such. Mostly just some old dust/mold on the main board inputs. Cleaned most of it away.
I did some probing around and from what i can find NRESET is high (5 V). The NRST on Q6 is however only 1.4 V, i am not sure if this is how it should be. I figured it should be either 5 V or 0 V, but i do not understand the circuit well enough yet... The clocks are looking ok.
Probing memory activity on the CPU and bus activity on the VGA controller it seems like all activity stops around 800 ms after power on. This seems a bit odd. If i short SW1, there is about 100 ms with activity and then nothing. I the stop in memory and bus activity seem to coincide with the output of the reset circuit IC86 going high.
My problem is that the only schematics that i have of the CPU and VGA if from the service manual for the waverunner 1. It looks like there have been changes on this exact circuit on the waverunner 2 series.
Q10 does not connect directly to the output of the IC86 on my scope. There is at least one extra transistor that switches the connection between IC86 and Q10. I have however not been able to find anything that connects to the base of this transistor.
I think i will do some more digging tomorrow, got a feeling that i should investigate the NRST signal and IC86 a bit more.
-
FWIW & if it can be of any help it should be possible to boot the scope from an adequate SRAM card containing the firmware, useful to recover from a bad flash or when in doubt on the onboard flash IC's themselves.
This was at least possible with the 93xx and LC series scopes, unfortunately there is no documentation around on that feature for the LT series.
I assume it involves jumpering somewhere on P2 and that the last four lines in the table at the bottom of sheet "BUSCONTROL (33)" on the WR1 service manual also relates to that although I can't see how to apply it to the WR2.
I'd have gone out experimenting on my LT but I'm stuck for a way of getting the firmware image onto a SRAM card.
Concerning NRST: 1.4V sounds OK, R17 and R18 voltage divider same on WR1 and WR2.
NRST goes low either on a SW1 press or driven by Q6. I assume Q6 is part of the front panel key combination software reset.
-
Ok. Thanks! I have looked into the possibility of flashing new firmware, but only have a CF adapter in my PCMCIA slot right now. Would it work with a CF card ? I also have no idea on where i should start looking to get a copy of the firmware even if i could figure out how.
I cannot find any bus activity out from the VGA controller, but i guess that might also be due to none being sent from the CPU. Nothing on pin 53 (activity indicator) of the VGA controller either.
On the CPU there is only about 100 ms of activity between the RAM and the CPU after every reset. If i disconnect the RAM, then there is no activity at all and the scope does not even reach the stage where the standby led turns on. Disconnecting the ribbon cables to the main board also have no effect. Still only 100 ms of activity when reset is triggered.
Based on everything so far, it looks to me like a problem either with the RAM or the CHIPS VGA controller. I think i will try to get hold of a new stick of RAM and possibly also order a new VGA controller if i can find some cheap on AliExpress or something.
-
For the firmware and other advice subscribe to the LeCroy owners on groups.io
If you don't want to bother you can get the firmware here until I need to free some storage space:
http://eninz.free.fr/wr2-9.3.0/WR2_(Blue_Cabinet)_ver_9.3.0.zip (http://eninz.free.fr/wr2-9.3.0/WR2_(Blue_Cabinet)_ver_9.3.0.zip)
I don't think you can update or rescue boot from CF.
You'll need a minimum of 2MB PCMCIA SRAM which are expensive and not that easy to come by, after that you still don't know the jumper config to start from the card.
Definitely give changing the SDRAM a try first that should work-out cheap.
-
Thanks for all your valuable help. I am happy to i form that i finally found the fault. It looks like it was just a bad memory stick. I hope that the information in this thread can be of help to anyone else that have similar issues.
My only issue now is that the trigger calibration fails and that the scope freezes up and displays an error message if I hit the panel button. Not a huge issue for me at least. Will do some more digging to see if I can find out. Does this scope have a self test menu or similar that I can enter to test all systems ?
-
Lucky for the SDRAM.
Well... a working trigger is a useful feature on any scope, does it just fail calibration or not trigger at all?
The maintenance menu and error log is accessed by exiting any menu by pressing the return key a number of times.
Pressing and holding 3rd menu key, pressing and holding 4th menu key, pressing and holding 5th menu key, releasing 3rd then 4th then 5th.
-
Thank you! That menu looks very helpful although a bit overwhelming. Yes, a working trigger is pretty essential :)
I enabled calibration logs and found that all channels fails the trigger test on both positive and negative, AC and DC coupled.
From some further testing i have confirmed that the trigger function is entirely broken. It is not possible to get any of the channels to trigger on anything.
I guess i will have to do some more probing to figure out what is wrong with the trigger circuit.
Will start with checking all the power rails for MST429A, the _RESET and the different clocks. The MST429A has a -2.3 V rail that i have not probed before. Luckily the trigger circuit looks pretty well documented in the service manual, i have a decent hope of finding the problem. I really hope that the MST429A is not broken, in that case i might be at the end of the road...
-
I've had the problem of the -4.5V (VEE) supply to the MST412 ASIC intermittent due to a bad Melf diode on a LeCroy 9354, the diode has been changed to a different setup on the LT acquisition board but I'd definitely check that top of the list. (Including the filter 11FL1)
-
Sorry, wrong SM... I was referring to WR1, didn't remember there was the schematic for that section in the WR2 SM.
Anyway the same applies... Check common power supplies involved, right up to the ASIC if possible. Then you get the specific ones VEE (-4.5Vmst), -2.3V Vtt(ECL), I think Vbb is generated by the MST429A itself.
If you must probe around the acquisition board make sure you get some airflow on it if running more than a few minutes.
-
The voltage rails looks to be ok, i have however found one potential issue. I cannot seem to pick up the 125 MHz clock. I am able to pick up the 500 MHz clock.
Probing on 12Q5 collector i only pick up -3.4 VDC. Probing on pin 7 on 12IC5 i only pick up -1.75 VDC. No clock signal that i can pick up at least, just 10-20 mV noise.
Probing the 500 MHz clock going into 12IC5, i am able to pick up a weak (150 mv pk-pk) signal with a 1.3 V DC offset. The reason for the weak signal is however likely due to the fact that i only have a 350 MHz bandwidth scope available to probe around with right now. I don't really know what signal level to expect for the 500 MHz clock.
I will check at work if i there is a faster scope that i can borrow, but for now i am assuming that the 500 MHz clock is OK. It does however looks to me like the 125 MHz clock is mostly used for calibration and perhaps the adjustable calibration output. From what i can see, a bad/missing 125 MHz clock should not result on trigging functionality not working ? Except if the scope disables all trigging if the trigger calibration fails, that seem however seem a bit far fetched.
-
Grabbed a better scope from work to check the 500 MHz clock. Unfortunately I do not have any better probes on hand than the LeCroy PP006 500MHz probes that comes with the LT264. I have attached a plot of the 500 MHz clock. It does seem a bit too low to me. And that might also be the reason for the 125 MHz clock also not being present. Hope that it is one of the resistors or capacitors on the 500 MHz clock that has failed short. If not, then I guess the output driver for 500 MHz clock on the 14IC2 (MCG426) could be bad. Let's hope not.
-
I'd call that good, you're really on the edge of what you can probe that way...
For comparison here's what I got with the same probes you used and a DDA-120 as display while looking at the CKL/ _CKL inputs to the MST412 I was on about previously.
I had doubts at the time too but it turned out the clock input was good.
-
Ok, thanks. That makes sense. Did some reading and the clock is referenced to VBB (between -1.35 and -1.25 V in this case), i guess that explains the DC offset of around -1.3 V. It might be that the 500 MHz is ok yes. The divider circuit does however not produce a 125 MHz CLK, at least i am not able to capture one with my setup. I might order a MC10EL33 in case that i need to replace it.
-
For probing at the fast ECL signals you'd probably be better off with a piece of 50 Ohm coax and 50 Ohm termination on the probing scope.
I'm not sure about this but if the "125MHz CLK" is not active all the time I'd expect it to be at least during power on reset and during an automatic recalibration if that's enabled.
Do you ever see "INTLV CAL" change state during POR?
-
INTLV CAL goes high for about 200 ms on startup before going low again (+700 mV, - 700 mV). It stays low during the self test and initial calibration.
That being said, i know nothing about ECL circuits. For all i know - 700 mV is high and the 200 ms of +700 mV is just some power on thing. From what i understand in the datasheet for the MC10EL33, a high reset pin (INTLV) will cause the chip to not output a clock. I am however a bit confused on how to understand the logic high and low signals on a scope with ECL signals. I guess the signaling is current based and therefore a not so simple to measure in terms of absolute voltage levels.
It is very easy to pick up the 500 MHz, at least when i borrowed a faster scope. The 125 MHz i am not able to pick up at all.
It might be wrong, but the way i see it right now one of these three is true:
1. 500 MHz is too weak to for the MC10EL33 to pick up (maybe doubtful seeing that i can pick op the signal and it looks clean. There are also no oblivious shorts, it charges up to between 5-20 kOhm when measuring the resistance to ground on both the 500 MHz and the 125 MHz)
2. I am misinterpreting the INTLV/Reset signal (- 700 mV is high) and therefore the chip is set to not output the 125 MHz signal
3. The MC10EL33 is bad and needs replacing, i have ordered one just in case.
I guess it is also another option that i am completely wrong and that the missing 125 MHz clock has nothing to do with all this. After all, it does not look like the 125 MHz clock is used to anything other than the generation of the calibration signal for the self test. I would assume that it would fail on other things than the trigger calibration if it does not have any calibration signal at all ?
-
I don't see why you'd see any negative values for "INTLV CAL", something is wrong. Is your scope DC coupled and reference for probing GND?
I'd expect Low as around 0.3V, High around 3.8V (being pessimistic) and high impedance to rest at 1.8V, that translates to around -2.8V, -1V and -1.3V for high impedance at the MC10EL33 reset input which just about fits the logic requirements.
MC10EL33 would be reset for most positive values (>-1V on pin 1).
Have you tried setting your scope to single trigger on the "125MHz CLK" during sartup or tried looking for one of the "CHx CAL SIG"?
-
Yup. DC coupled and ground reference.
Tried setting the input impedance to 50 Ohm, then the INTLV CAL is stuck at 0 V. Never goes high (if that is >0 V)
Tried to single trigger on both 125 MHz and the channel cal signals and I get nothing. Measuring the 125 MHz I get a steady DC voltage of about 3.3 V.
In summary, I measure the INTLV CAL to be stuck at 0 V when probing with 50 Ohm input and around -700 mV when probing with high impedance. Is that sufficient to set Reset on the MC10EL33 and thereby not output any 125 MHz ?
-
My comment on probing at 50 ohm coupling was only for the fast ECL outputs.
Regarding "INTLV CAL" (16IC6 pin 7): Don't confuse it with the pin 1 input to 12IC5.
Considering values of resistors 12R82, 12R83, 16R53 on schematic it shouldn't be possible to have a negative voltage at "INTLV CAL". 16IC6 is supplied between +5V and GND so the only case I can imagine to have a negative voltage there is 16IC6 being high impedance or having a defective open output and 16R53 being open. Although that situation should lead to 12IC5 outputs being active and not in a reset state...
Per datasheet 12IC5 divider should be active when its pin 1 is in the -1.5 to -5V range and that would only happen when 16IC6 pin 7 is held low.
If schematic is correct and for the sake of checking the functionality of 12IC5 without desoldering it, you should be able to force it active either by removing 12R82 or "shorting" 12R83 (parallel a low value resistor).
-
Yes, that makes sense, my mistake. Probing on the actual "INTLV CAL" signal gives me a steady +5 V. This is constant even during power on self test. It looks like reset is indeed high.
Shorting out 12R83 does produce a 125 MHz clock, i guess proving both that the 500 MHz is ok and that MC10EL33 is working correctly.
I will have to do some more investigation on what should make the "INTLV CAL" go low. I would assume that "INTLV CAL" should go low during auto calibration, but i assumed that auto calibration runs during power on self tests ? I also tried to do a test from the service menu (picture attached) and at no point during this test did "INTLV CAL" go low.
That being said, i still don't know what the circuit connected to the 125 MHz actually do, other than supply the clock for generating CHx CAL SIG. Or in other terms, i still do not know what CHx CAL SIG does. I assume that it is a for a calibration routine, but all other items than the trigger passes the self test. Only thing i know is that it seems like CHx CAL SIG is not being generated during power on self test and the self test in the service menu. Although, for all i know that might be normal. I still does not know what triggers the scope to enable the CHx CAL SIG or INTLV CAL...
-
Sheet 17 (page 149/155) shows the 16bit DAC that is responsible for generating the variable control voltages for most of the frontend and trigger control.
DAC's buffered output "DCG CHARGE" is then sent to be MUXed out to the various places it's needed, this is done by three 4051's on sheet 18.
One of these MUXed outputs becomes "INT CAL REF", which after going trough 16IC1 and 16IC2 becomes "INT CAL".
"INT CAL" is sent to every frontend where it can be switched in place of the BNC input to calibrate the attenuator, BWL, and amplifier sections.
(I'm not quite sure why they have "INT CAL" and "CAL OUT LVL" on 16IC4 unless that's a diagnostics function available on the front panel CAL OUT.
Anyway that's how it could pass other stuff and fail trigger, the main calibration signal is "INT CAL"...
On sheet 18, the DDS (16IC8) should be able to generate an output frequency up to 60MHz, because this goes through an ECL buffer then series capacitor I assume the "CHx CAL SIG" that go to frontend HFE428's are only used for timing and not amplitude. If it is for calibrating the interleave when combining channels as "INTLV CAL" name may suggest I'm not sure how it would be used on the LT264 as it doesn't have the ability to combine channels anyway. However not having RTFM I'm a little lost here.
Back to the trigger:
If you're not getting any change on "INTLV CAL" at any time, check activity and signal integrity on 16IC6 pins 11, 12, 14. Better still to scope all three at the same time and you may have a chance of reading if the LT264 actually loads the data to enable "INTLV CAL".
Also check 12Q2A, 12Q2B if that's fried it could be holding the line high.
Another suspect if you have no trigger at all including on EXT Trig. is 18IC3 (page 150/155) as this MUX is responsible for all the MST reference voltages. Analog MUXes are known to fail once in a while. Sheet 12 gives the voltage ranges you should get after the buffer OP-AMP's (+0.56V to -3.07V).
Following that route it would also be interesting to check if you can get any output on the front panel CAL OUT ring. If you do have some output try changing the amplitude setting in the scope's menu that should give you some clue on whether 18IC3 is completely dead or not.
-
Correction: The voltages to expect after 18IC3's buffer Op-AMP's are in the +4V to -4V range the ones stated previously are at the MST429A after the level shifting resistors.
I suppose you do get inputs to MST429A "CHx TRIG" and "CHx VAL" when trigger is set correctly.
Have you looked at MST429A pins 98&99, 93 for an output?
-
Again, thank you for all the valuable information. I am working though your suggestions one by one and trying to test as much as possible of the different options.
From what i am seeing so far, 16IC6 looks to have sensible signal levels and activity both in and out, just not on INTLV CAL. I have yet to try to decode the if INTLV CAL should be on. This is on my todo list as it will require me to perhaps solder in some test leads, i am a bit short on hands :) It might well be that INTLV CAL is not used at all on this scope during normal operation for all i know.
Yes, ext trigger is also dead. I have checked 18IC3 and it looks good to me with values in the specified range.
As for the front panel calibration out, everything looks to be working correctly with all different frequencies and amplitudes.
I am having a hard time picking up any signals on the CHx VAL lines, even when i feed in the calibration signal. I have not been able to pick up anything at all for now. I am finding this a bit strange, but not quite sure what signal levels to expect. As for the CHx TRIG, i am able to pick up signal being toggled here during the auto calibration, but nothing other than that.
I will tru to probe the MST429A pins 93, 98 and 99 as you suggest next.
I also found what could look like a leaky capacitor 21C46, picture attached. Do not necessarily think that this is part of the problem as this is one of several on the -5V rail. I have measured this rail earlier to be correct, at least the voltage is. I will however try to replace it at some point.
-
Your capacitor looks more like having solder flux residue than a leak.
I'm puzzled to why you'd see "CHx TRIG" change during auto cal. but not during normal use... :-//
Maybe It's worth starting a little closer to the signal input?
Let's assume you're troubleshooting on CH1, input a known signal to your LT264 on CH1 close to full vertical scale and confirm that the displayed waveform has correct amplitude with another scope.
Set the trigger source to CH1 and level so it's around median of your input waveform also set trigger to AUTO.
If I read SM correctly you should have in the range of 250mV peak at 2IC4 (HTR420) pins 29 and 30 (500mV differential) at this point it is still analog.
Can you confirm that you get no toggling on 2IC4 pins 11, 12, 14, 15 outputs?
-
The signal level going into HTR420 (pin 29/30) looks correct. I get 500 mV peak to peak on each of them with the cal signal (verified that this is correct with a separate scope) and setting the range so that i maxes it out.
Pin 11/12 have a steady -2.2 V DC. At boot during calibration it goes from 0 V to -1 V and then to -2 V and get stuck there.
On pin 14/15, the signal is at a steady -1.85 V DC. At boot during calibration it does switch a couple times between -1 V and -2 V.
If this is not how HTR420 should behave i am a bit surprised that all of the HTR420's act the same. It is not that many things that are common between them.
-
Measured inputs sound good, I'm assuming that the trig outputs are good enough too as their levels do cross typical NECL input thresholds.
Although I'm not sure what the serial bus on the HTR420 is used for (I don't know if it can inhibit trigger outputs preventing a new trigger sequence if the trigger signal is lost further down the way) it is one of the things that's in common with all HTR420's but also with other functions like HFE428's and FE relay switching.
I'm also thinking the same, odd that all inputs should fail trigger when there is not that much in common.
Next points I'd test are:
+/-5V supply rails at the HTR420 pins.
DC voltage variation at pin 32. Should be in the +/-680mV range (possibly as low as +/-500mV) and vary while adjusting the trigger level for the selected channel.
Consistency of "FE SRCLK", "FE SRDAT","CHx FE LD","CHx HTR LD" lines.
I wouldn't bet on the first three as they would likely also cause other problems than just the trigger. The "CHx HTR LD" however are dedicated to the HTR420's only.
CH1-4 have a common origin in 20IC4, unfortunately "EXT HTR LD" comes from a different 3 to 8 line decoder 20IC6.
However a stuck-low output on 20IC5 pin 15 would cause all five "xxx HTR LD" to not change, it would also cause "CAL LD" to not change which would in turn cause "INTLV CAL" to not change... So definitely worth keeping an eye on 20IC5 it all seems to point that way!
-
I really appreciate the input on this, thanks again.
Unfortunetly 20IC5 pin 15 has activity on it and there is more activity when i change around the trigger and such. It looks to me like it is working as intended, really good idea though.
I am having a hard time imagining what could be wrong at this point. Everything is measuring ok. The +/- 5 V supply rails for the HTR420 looks ok. I will try to measure the rest of the FE SRCLK and FE SRDAT lines tomorrow.
Pin 32 is at a DC level that have a range of 300 mV to 600 mV, depending on the trigger level. There are however quite a bit of noise. If i am set to single trigger or normal, then the noise is about 100 mV or so. In auto, it is up to 500 mV or more. There are also a noise spike that corresponds to 32 kHz (300-400 mV spike every 31 us or so). Don't know if this is to be expected, but that is at least what i measure.
I have also gone over the board today on both sides with a thermal camera to look for failed capacitors and such without any obvious things showing up.
I think i will order some more memory to try out as well, it seems that the memory i put in is only rated at PC100 and not PC133.
One other thing that i want to try out is to reflash the firmware on it. Would it be possible for the memory issues to cause some data corruption on the firmware ? As mentioned earlier, the scope crashes every time i hit the "pannels" button as well. Seems like it cannot reach some address or something. Locks up and requires a reset.
I was hoping to reflash it today using floppy disks, but unfortunately it looks like the floppy drive is bad also (seems like a common issue). Will try to get hold of a GPIB cable from work tomorrow as well.
-
I'd definitely expect HTR420 pin 32 to be able to cross 0V when adjusting the trig. level (unless I'm confused and "CHx TRIG LVL2" is the main trigger and "CHx TRIG LVL1" is unimportant for basic triggering, still I wouldn't expect all that noise).
If you look back at sheet 18 (DC GENE) the "DCG CHARGE" is stated as having a +/-4V excursion, that is muxed by a couple of 4051's then buffered by TL084's and finally brought to adequate levels with a resistive divider before the MTR420's.
100mV noise (be it peak, p-p or RMS...) sounds atrociously much too much. Maybe it's worth working through the path from "DCG CHARGE" to see if something is wrong on the way.
Once you've checked that and the three serial bus lines to the MTR420, I'm out of ideas.
The firmware refresh could be a good idea, hard to decide if a corrupt/wrong firmware is causing the problem or if a hardware fault is causing the software to hang...
I don't know if it is possible to update a WR2 with a WR1 firmware and get into strange situations either...
ISTR there was a CRC check tool on the groups.io "LecroyOwnersGroup"although I don't remember if it was used for all LeCroy DSO firmwares.
-
HTR420 pin 32 will cross 0 V and turn negative if i increase the trigger no near the maximum level. Then it suddenly flips and turns negative. Maximum negative trigger level is still well in the positive numbers.
As for the noise, i think that some of it might be a case of bad probing. I seem to pick up a lot of high frequency stuff in the ground lead for the probe. Tried again with a probe ground spring and got much lower values in the 50 mV range, probably some of it is still radiated noise. With just a ground lead on the probe, i get up to 700 mV (about 50 mV or so with the ground spring on the probe) noise spikes on the +/- 5 V rail. Unfortunately it is a bit hard to get to some of the probe points with the ground spring setup.
I am getting close to giving up on it, but will try to reload new firmware on it and also try a new stick of RAM once that arrives. This is a bit of a mystery to me as well. The time i have spent on this i probably would be better of just throwing it away and by a cheap rigol scope or something. I do appreciate how snappy it feels compared to some of the Tektronix scopes i am used to from work and it seems to have much more features as standard as well.
-
Another ting i find a bit odd after some additional probing is that pin 1 on 2IC5 (3IC5, 4IC5 and 5IC5) never goes high, at least duing auto cal at boot. I guess that is intended behavior ? I guess that it makes sense, seeing that it seems like this cal signal is not used, at least not for power on self test or auto cal.
-
HTR420 pin 32 will cross 0 V and turn negative if i increase the trigger no near the maximum level. Then it suddenly flips and turns negative. Maximum negative trigger level is still well in the positive numbers.
As for the noise, i think that some of it might be a case of bad probing. I seem to pick up a lot of high frequency stuff in the ground lead for the probe. Tried again with a probe ground spring and got much lower values in the 50 mV range, probably some of it is still radiated noise. With just a ground lead on the probe, i get up to 700 mV (about 50 mV or so with the ground spring on the probe) noise spikes on the +/- 5 V rail. Unfortunately it is a bit hard to get to some of the probe points with the ground spring setup.
I am getting close to giving up on it, but will try to reload new firmware on it and also try a new stick of RAM once that arrives. This is a bit of a mystery to me as well. The time i have spent on this i probably would be better of just throwing it away and by a cheap rigol scope or something. I do appreciate how snappy it feels compared to some of the Tektronix scopes i am used to from work and it seems to have much more features as standard as well.
Did you mean:
"HTR420 pin 32 will cross 0 V and turn negative if i increase the trigger no near the maximum level. Then it suddenly flips and turns positive. Maximum negative trigger level is still well in the positive numbers."?
In any case I can't make sense of this, can't see how things should work that way...
Hopefully the firmware refresh will get it back on track, they are nice scopes when they work, plenty of features even without software options enabled.
Another ting i find a bit odd after some additional probing is that pin 1 on 2IC5 (3IC5, 4IC5 and 5IC5) never goes high, at least duing auto cal at boot. I guess that is intended behavior ? I guess that it makes sense, seeing that it seems like this cal signal is not used, at least not for power on self test or auto cal.
It at least fits with "INTLV CAL" not being active either, whether it is supposed to be that way or not is still uncertain...
-
No. Maximum negative trigger (- 2.5 V at 0.5 V per division) is at about 324 mV, then it increases with increasing trigger level until it reaches 600 mV at 2 V trigger level (0.5 V/div). At 2 V trigger level, pin 32 will suddenly flip to - 735 mV and increasing the trigger level to 2.5 V will increase the voltage at pin 32 to - 700 mV
Sounds a bit stange to me, almost as if there is an offset error somewhere for the trigger level ?
Did the reflash of the firmware, unfortunately it did not fix the issue. I am waiting for a new PC133 stick of ram to replace the PC100 that i use now, maybe that can make an impact, but i have my doubts.
Another thing that have bothered me is the fact that if i hit stop on the trigger control, time controls do not do anything. I can adjust V/div, but adjusting s/div does nothing. I can see the s/div changing, but the graph is not changing Is that normal behavior on LeCroy scopes ? The scope would be almost usable if i could just trigger manually by hitting stop and using the relatively decent memory depth to zoom into what i am interested in...
-
That's clearer thanks.
There is a problem to solve.
I've opened my (working) LT and checked a few points where it was convenient to probe.
Set to 1ms/div, 1V/div, auto, edge, DC, pos.
Measured "CH1 TRIG LVL1" at 2R105 coming from 18IC10 before resistive divider.
Trace centred, trig level 0V. : +66mV
Trace centred, trig level 4V (end of graticule): +1.55V
Trace centred, trig level 5V (maxed out above graticule): +1.92V
Trace centred, trig level -4V (end of graticule): -1.41V
Trace centred, trig level -5V (maxed out below graticule): -1.78V
Trace offset +2V, trig level 0V: +810mV
Trace offset -2V, trig level 0V: -670mV
There's no sudden jumping around, unless the auto recalibration kicks in, no nutty flipping negative when adjusting upwards...
Second reference voltage "CH1 TRIG LVL2" can be varied a little (measured at 2R107) by setting trigger to "window" and adjusting window size.
On the output side of things:
Idle (untriggered) voltages are -0.92 and -1.7V although I didn't check which is _OUT1 v.s. OUT1.
"CH1 TRIG" changes state when the trigger level is adjusted to a level that allows to trigger.
"CH1 VAL" will kick-in as well as "CH1 TRIG" when scope triggers using the window trigger.
Try working back from "CH1 TRIG LVL1" to find where things go wrong. I'd give the whole DC GENE bunch a reflow starting with 17IC1.
If you don't get to the bottom of things you should be able to use your trigger by removing the erroneous "CH1 TRIG LVL1" and feeding in your own variable reference, of course the scope will still fail self test...
Supposing that the panel menu hang is not related to the trigger problem, I'd check if it will also crash when accessing the "mass storage utilities" menu.
If it doesn't, look towards RTC and SRAM on CPU board, re-check for oxidization there at component pins. Maybe worth swapping the SRAM I.C.
If you stop the acquisition you can't expand the capture directly by changing time/div, you have to go through the magnifying glass function...
Edit: I forgot to mention that there's no obvious noise on "CH1 TRIG LVL1".
-
Wow! Thank you. This is immensely helpful.
To put tings into perspective, i measure about 3.1 V at 2R105 when the trigger level is set to 0 V with the same settings. From there in keeps increasing as i increase the trigger level before it wraps around to negative.
Trace centred, trig level 5 V: -3.8 V
Trace centred, trig level 4.1 V: -3.9 V
Trace centred, trig level 4 V: +3.8 V
Trace centred, trig level 0 V: +3.1 V
Trace centred, trig level -4 V: +2.4 V
Trace centred, trig level -5 V: +2.2 V
There is certainly something wrong here. Based on this, it makes sense that the scope does not trigger. The measured trigger levels are set so high that even at -5 V set trigger level, i would need a voltage probably in the 6-8 V range. That being said, i was still not able to get it to trigger even if i applied a 15 V signal. The front end would probably start clipping at those kind of levels i would guess. Maybe i have so much offset in the trigger level that the front end start clipping the signal before i ever reach the required signal level to make the scope trigger.
I guess that i need to digg more into the DC GENE circuitry and the output from 17IC2. Perhaps identify the part of the muxed signal that corresponds to the trigger level(s) and check if it also have a significant offset in the output from 17IC2. 20IC1 could also be interesting as it adjusts the DCG offset voltage and DCG gain. In theory it should be possible to adjust the DCG offset in software (if that possibility exists in the service menus), but i will not start messing with that yet as the potensial offset error is so large that i have i hard time imagining that it is just a matter of calibrating the DCG offset.
I guess that i would need to decode the signal going into 17IC2 in order to actually check and if needed adjust the offset as it would be hard for me to know if there is an actual offset error from the IC without knowing what voltage is commanded over the digital input. That being said, i could probably assume that the digital bit is correct for now.
I think i have some work ahead of me :) Again, i am really grateful for these measurements. It seems like this could be a major step into figuring this out. Once i have gone over all the remaining tings that i want to probe on these chips, i will do a quick reflow of them, just in case.
-
I have attached pictures of what becomes "DCG CHARGE" with trigger level:
-5 V
0 V
4 V
5 V
This is on 17IC2 pin 3 (VOUT). This is a +/- 10 V signal that goes into a inverting op-amp with a gain of 0.4.
I have made a red circle around what i think is the signal that corresponds to the trigger level ch 4 (i have used ch 4 as it is more easily available for me to probe).
As shown in the pictures, this strange trigger level with the large offset comes directly from 17IC2.
It seems like a bit of a hassle to probe the 16 bits that goes into 17IC2 (DAC712UK) to be able to compare the set voltage with the measured voltage out of 17IC2.
Without probing the digital signal going in to 17IC2, i find it a bit hard to tell if the offset on 17IC2 is wrong or not. Since it seems like the signal will just wrap to a positive value (as measured before the inverting op-amp) if the signal exceeds - 10 V. The signal seems to be centered well around 0 V at startup at least (offset of around 20 mV or less).
As strange as it sounds, it almost looks to me like the offset is in the digital signal as well. I guess i need to try to decode the DCG telegrams going into 17IC1 to be sure.
-
If you have selected the right portion of the "scanning" DAC's output sequence it does indeed look as if the problem is already there that far back.
Good idea to check on "DCG OFFSET", easier to get out of the way before moving on to the digital side of the DC GENE section.
I don't know about the 17IC3, 17IC4, 17IC5, 17IC6, 17IC7B portion: It is used to detect the probe identification resistor on the sense ring and channel 50ohm input overload but could also be used to setup "DCG OFFSET" by selecting one of the grounded inputs on 17IC3 or 17IC4 then searching the comparator switching point that is then reported back to the 17IC1 microcontroller. Keep those two 4051's in mind if you do find that "DCG OFFSET" throws the trigger control voltages around.
I also agree that it's tempting to come to the conclusion that the digital side is working if no other scope functions depending on the scanning DAC are experiencing erratic behaviour.
The digital side could be electrically sound but have the wrong data sent to a DAC because of a corrupt EEPROM.
-
Here's a .zip full of screen dumps from my LT354 if that can be of help investigating.
If you spot trigger related values that are way-off you could try modifying one channel manually for the sake of it. I'd advise to save a copy of the original cal data first.
Have you tried hitting the "UPDATE ALL ITEMS" contextual menu when you're in the "CAL ITEMS" menu? Does that result in a trigger error?
-
Thanks again. These could prove very valuable for further analysis. I just did the same measurements and are in the process of comparing the values. I will probably spend some time converting everything from HEX and putting it in a excel sheet together with my measurements to get a clearer picture of everything.
Yes, i have run the "UPDATE ALL ITEMS" command in the "CAL ITEMS" menu. I fails on the trigger level just as it switches over to the trigger test. Then spends some time with testing trigger before it beeps and say NG on the trigger.
I does look like the HTR420 trigger test is able to pick up anything ? Results in a gain number of 6000. Slope+ is at 71535 and slope- is 6000. Same for all channels (some have slightly lower gain).
There also looks to be something strange with the measured trigger hysteresis. In the hardware register "Trigger Hyst B" and "Trigger Hyst D" have a "voltage at output" value at just over 7 V. "Trigger Hyst A" and C is at some impossibly large number. I have measured the actual signal to be about 2 V at 5C94 for all channels. That sounds about right, based on the register value for the signal going into the DACs. On pin 34 of the HTR420 i measure about 2.6 V for all channels.
I did do some more digging on the probe identification circuitry i actually fond a bad contact between the probe detection ring and the main board. There is a small gold spring that touches a pad on the board of all the different channels. Two of the channels had some soot on the pad, creating a bad contact. That should however only affect the probe detection functionality from what i can tell. I cleaned the pads with some contact cleaner and now all the channels are able to detect probes properly. I measure 4 V on the probe rings, guess that should be about right based on the schematics.
I have also done some calculations on the signal controlling gain and offset for 17IC2. The values i measure going into 17IC2 matches the calculated values based on the values from the registry regarding the digital signal going into the DAC (20IC1). I belive now that gain and offset is set correctly. It is the digital signal going into 17IC2 that is strange.
-
Just a thought trying to link the two problems your scope has:
Of all the calibration values I have no idea which are factory set with "LeCalSoft", and what is able to update automatically during normal use to compensate for temperature drift.
I don't know either where that data would be stored, if it is split into several locations or redundant.
There is an EEPROM on the acquisition board, and SRAM enabled through the RTC on CPU board.
If you suspect that the wrong data is being loaded to the scanning DAC (which is looking more and more likely), and considering that the scope hangs when you go to the "Panels" menu (which I can only imagine are saved to SRAM).
I've never heard of a LeCroy DSO go out of calibration because of a dead RTC battery which also maintains SRAM data so the factory calibration won't be stored there.
It could be that on boot the factory cal. is read from acquisition EEPROM, copied to SRAM and used from there.
Maybe a bit far fetched but not absolutely impossible.
SRAM would also be a good suspect because the scope was "exposed to the elements" and as long as the battery was charged the SRAM was powered making humidity a more destructive hazard.
-
The whole thing with calibration seems a bit strange on these scopes. I don't know how much factory calibration there really is on these. It almost seems like it measures and calibrate most if not all of the cal parameters itself. Looking at the replacement procedure on the main board, it just states that following a main board swap, one need to run the "Measure board items" and the "measure overload" functions and save them to EEPROM. There are mentions of manual main board adjustments, but i assume that is the adjustment of the potmeters on the board and such.
The "measure overload" function is a bit interesting. I don't know what it does, but i guess it is not too far fetched to think that it might check some ranges and offsets.
Right now i have the scope in 100 prices as i thought it would be good to have a quick look at the power supply as well, but once i have put it back together i will definitively try to run the "measure overload" function and try to investigate what it does and if it could potentially change anything about the big trigger level offset error that i have. Maybe the information that "measure overload" acquires is corrupt or missing. I see from the cal registry that it does look like all the channels are missing hysteresis data.
One thing that i do however find interesting is that during the initial calibration at boot, it looks like there is no offset to the trigger level control signals. It might also be that the offset is a result of the failed trigger calibration, but i have to think about this a bit more if this is a possibility. On second thought, it would be interesting to scope both the signal into HTR420 and the trigger level at the same time during calibration. Signal levels should be fairly similar and that should allow for me to confirm if the HTR420 should trigger or not. Another thing, is it to be expected that the voltage going into pin 34 is as high as 2.5-2.6 V ? I am a bit confused as all the other analog signals going into the HTR420 is scaled down to +/- 600 mV range. As someone that are more used to working with kV's and MW's, i am a bit behind on my circuit analysis theory :) I cannot seem to understand how i can judge how the voltage is scaled by 2R110, 2R109 and 2Q12. More used to transistors used as switches and not as amps :)
I have looked at the circuitry around the RTC circuit and what i assume is the SRAM. It looks to be in good condition at least visually. Even though the scope was exposed to the element (inside a shipping container), it looks to be in very good condition. The only thing that i have found so far that is a bit concerning is that some of the exposed gold/copper traces have partly corroded away some places. It does however looks like there is no broken traces, it is mostly the exposed shielding traces that are corroded. Other than that, it is things like a little-bit of dust/mold in the fan inlets/outlets. That is however mostly sorted at this point.
-
Did you get any further on this?
I'm working on a long repair project a Tek TDS620 that has suffered from the usual SMD electrolytic havok and experiencing similar issues. Some of the FE analog control voltages are missing, I've double-checked the DAC is working, tested track continuity between there and analog MUXes, replaced a couple in doubt, no better so I'm pretty sure the corresponding data is lost before on the CPU board, just finding where is a nightmare...
As for your LT's hysteresis control on pin34 of HTR420 I'm not sure what to expect either the output from DAC is in the 0-4V range, I'd expect the scope calibrates the hysteresis control level itself, anyway as long as it's mot trying to exceed the +5V to the HTR I'd call it good.
-
Sorry for the late response, been away on holiday. I am planning to start to work on the scope again in the coming weeks, will let you know if there is any updates in case it might be of interest to you or others having similar issues. I am still stuck with not being able to trigger and the strange offset on the trigger control voltage on the DAC output.
Tried to change the RAM sticks again and run the measure overload routines and similar without any improvement.
-
I have to say it would be very interesting to know the purpose of the 16IC8 and the whole circuitry creating "CHx CAL SIGNAL".
Have tried to trigger on these signals and i am not able to detect anything on either of these no matter what calibration sequence or "measure overload" sequence i am running from the service menu.
I mean, it surly have to have a function right ? Could it be that this circuitry is used for the trigger calibration, hysteresis and possibly the adjustment of the DCG trigger control voltage that are off ?
I feel a bit stuck so it is probably worth to check out further i guess.
-
I'll try to remember that and see if I can get a few captures next time I have an excuse to open my WR2.
No idea when that will happen, I'm not finding much time or motivation for electronics lately.
-
I think that i have found a situation where "INTLV CAL" goes low and i guess that 16IC8 AD9850 should be active.
In the trigger calibration/adjustment menu, there is a "ODP/EVT/HYST auto adjust" button. This starts calibration routine that activates the "INTLV CAL". This routine seems to get stuck for a while at "Incomplete RIS acquisition 1 of 50" before it continues and stop. When this routine runs, it seems like that the circuitry around 12IC5 is active and a 125 MHz signal goes into 16IC8 (AD9850), the signal is however very weak. See the attached scope pic. If i read the datasheet correctly, the signal going into the AD9850 should be a CMOS level signal and i would assume that a 4 V signal with a 50 mV 125 MHz signal would not be sufficient ? There is also so signal out of the AD9850. The voltage rails for the AD9850 and the reset looks to be correct so my theory is that either the 125MHz CLK signal is too low amplitude or the AD9850 is bad (or both).
My next plan to start to analyze the circuitry that i assume should convert the ECL signal from the MC10EL33 to a CMOS level signal. To be honest i find the circuit a bit intimidating, but i guess i can just start to try to measure diodes and shorts on the transistors, diodes and capacitors in the circuit. Hopefully one of these will have a clear indication of a fault. If not, i may try to change out the MC10EL33 since i already ordered a couple of those a while back in case they were bad.
-
I am struggling to find the component that cause the low signal level, but according to a application manual i found on the AD9850, the clock signal should be much higher.
"Proper clocking of the AD9850 occurs when the signal is at least 3 V p-p and centered at VDD/2.", https://www.analog.com/media/en/technical-documentation/application-notes/AN-419.pdf (https://www.analog.com/media/en/technical-documentation/application-notes/AN-419.pdf)
I will have to double check that i have not done something stupid like set a bandwidth limiter on or choose the wrong probes, but as long as i am probing this correctly and only getting like 50 mV amplitude on the 125 MHz signal, i think that i might be on to something that could possibly cause the issues that i am seeing.
-
Measuring the voltage out of the MC10EL33 i get a voltage swing of more than 3 V, as shown in the attached picture. The voltage at 12R78 is down to just 200 mV with a DC of about 4 V. Trying to measure the 12Q3, it measures a bit odd. Multimeter in diode test I measure a voltage drop as low as 0.15 V on one of the transistors between base and emitter. On the others, they are in the 0.8 V range. I am suspecting that 12Q3 or 12Q4 might be bad. I think I might desolder them to test them properly and change both if they still measure bad when they are off the board.
-
Base to emitter on 12Q4B measures in at O.L.
All the other transistors measures fine on 12Q4A and both on 12Q3. I have high hopes for this causing the issue.
Ordering a couple of these and changing them.
Will report back late next week if this solves the issue with the strange trigger control voltages.
-
I do not get it. I have now changed out 12Q4, 12Q3, 12IC5 (MC10EL33). I have desoldered and tested 12Q5 and 12D2. All voltage rails measure ok (12 V, -12 V, 5 V, -5 V and -2.3 V)
The circuit does still not work. Scope at 12R71 and 12R72 shows a strong (1.7 V 125 MHz) clock signal with a negative offset of about 1.3 V. The signal is completely gone at 12R78. I do not understand. I mean i have changed out or checked all active devices. The resistors are also measuring ok, at least as far as it is possible to measure in circuit.
Measure point at the base of 12Q4B (pin 4) show -7 V as it is indicated in the schematic that it should be. I guess the next step would be to desolder all the passive components and replace them as well. At least the ones that are difficult to measure in circuit. Had hoped that i did not have to do this :([attachimg=1]
Edit:
I measure 0 V on pin 2 and 6 of 12Q3. That surely cannot be correct ? Pin 1 and 3 measure 5 V. On pin 4 and 5 i measure the around -1.3 V signal with a strong 1V+ 125 MHz superimposed.
I also find it strange that i measure 12 V on pin 1 of 12Q4, i mean these are 11 V devices !? I could ofcource try to change out 12Q3, but i have already done it twice. However i might still try as sure does not look like it is working as it should...
-
Haha. Third time the charm i guess. Finally got the 125 MHz ECL to CMOS circuit working after a quick run with the iron to go over all the solder joints one more time :)
Bad news is that the trigger cal does still fail... At least there is a output signal on the AD9850 now during the "ODP/EVT/HYST auto adjust". About 200 mV, 32 MHz signal. The adjustments still seem to fail however for some strange reason. At least it is a step forward, the ECL to CMOS circuit was broken and the AD9850 did not output anything before. Even though it seems like it did not have any impact on the cal result for some reason.
Edit: Haha, if fixing the clock signal to the AD9850 seems to actually made things worse. Now offset01 cal and offset05 cal fails in addition to digitizer cal fail. Scope still seems to work though. I guess, there is more than one thing not working correctly and fixing one of them does not necessarily fix the others. I guess i should just continue trying to find other circuits not working as intended. Stating to get a bit feed up with this repair.