Upgrading Mainboard in Lecroy DDA-3000 (aka WavePro 7300a?)

[Converter]

Noise from filter capacitors through which large impulse currents flow is a fairly well-known phenomenon. As far as I understand, the reason for this is electrostriction. In this case, the pulse currents are most likely created in the capacitor by the switching transistors of the power factor corrector (PFC). Here, the duty ratio and the magnitude of the current in the pulse is strongly dependent on the mains voltage.
These switching transistors close the current through the PFC choke to accumulate energy in it at a certain point in time. At this time, there is a large current pulse through these capacitors (the same as in the release of this energy). The higher the supply voltage, the shorter these pulses will be, but the rate of current rise will be greater, because the inductor inductance is constant. Therefore, the voltage of the mains supply affects the level of noise.
Not all models of capacitors will show less noise after replacing the original ones, you need to select some of them.
Carefully, there is a limit on their height of these capacitors in the power supply case. My capacitors have a height of 18mm. It seems to be the limit.
Before this, it is also necessary to coat all chokes and transformers with a compound or varnish, since there is a high probability of noise from them as well.

[dietert1]

Isn't a capacitor that is working as a tweeter somewhat dissipative and prone to break down?
I just looked up some photos of our SDA 6020 power supply that i made while cleaning and it appears to me those three caps look more or less normal, yet with shadows or discolorations near their wires. What was the state of the old caps in your scope?

Regards, Dieter

[Converter]

Dieter,
They are the main current source for high-power PFC pulses (as I showed in the circuit above). I had to replace them many times in a torn state in the TDK-Lambda Vega series power supplies, where they fulfill the same role. Believe through them there is a huge pulse current.

My capacitors Vishay from the power source LeCroy do have some darkening on top of the film cover.
However, with careful testing using an LCR meter, they did not reveal any anomalies in electrical characteristics.
These Vishay F1773-447 no-boxed Metallized Polyester film capacitors are not a good choice for this place.

[dietert1]

At 230 VAC pulse currents are not as terrible and as an example 10 mOhms x 3 A x 3 A just gives 90 mW, so that won't melt a cap. Maybe there are really strong spikes from the PFC inductors parasitic capacitance. Pulse currents should be small during stand-by. Maybe without load the PFC controller doesn't run stable, but in some kind of burst mode.

At 230 VAC one might even turn off the PFC controller during standby. The PFC capacitor will then charge to 310 V instead of 400 V but that should work for stand-by.

PS: The Vishay F1773 datasheet, p6 says : "These  capacitors  are  not  intended  for  continuous  pulse  application.  For  these  situations  capacitors  of  the  AC  and  pulse  programs must be used."

Regards, Dieter

[Converter]

Hi Dieter,

As I understand it, you are now trying to calculate the average reactive current through a 0.47 μF capacitor at a frequency of 50 Hz?
But why no one sees that the load of these capacitors is an impulse current with a frequency of 40-100 kHz, which creates a PFC when it shorts the capacitors with a choke? What is the value of the impulse current there?
I will say it again: the voltage in the mains supply affects the rate of current rise in this choke and the duty cycle. Shorter pulses can create many harmonics and acoustic noise.

Sergey.

[dietert1]

No, no, at about 500 W the PFC choke will carry about 1.25 A  into the 400 V storage, right? Due to modulation the effective current may be about  3 A. The 50 Hz current into the first two filter caps is only about 0,034 A(rms) each at 230 V.

The 150 V/usec rating of those 470nF Vishay caps translates into about 70 A current pulse. It takes that current as a filter cap, but not in a continuous pulse application.

Regards, Dieter

[Converter]

Due to modulation the effective current may be about  3 A.

I think that much more. You do not take into account that they are powered not by constant voltage, but by the half-wave of the mains with a peak in the middle. We also do not know the duty cycle of these pulses. In the PFC of such power (600-700W), as a rule, you will see 3-4 MOSFET connected in parallel with 25A current limit for each (this is a necessary measure). At the same time, the traditional PWM converter, which follows after it, you will have a usual half-bridge with MOSFET 10-15A in each arm. This indicates a very different mode of operation of these cascades and the inapplicability of your simplified evaluation.

We must calculate the rate of current rise through the capacitor (which is equal to the current through the choke, since they circuited at the moment of opening the MOSFET) as: di/dt=U/L. I do not know the value of inductance and the pulse duration, so I will refrain from calculating the pulsed current through the inductance. Moreover, the control pulses due to a certain value of the time constant of negative feedback are non-constant in reality (overcompensation effect) at idle and difficult to predict.

I do not want to clutter up this thread with further discussion of the processes in power supply units, there are more suitable topics for this

[dietert1]

Sergey, PFC pulse currents are about 3x more when running at 80 VAC input (usually one of the design goals), while my estimate was for 230 VAC.

The outcome of this discussion with respect to those LeCroy scopes and alike:
- Hissing caps indicate problems. Those Vishay caps in the LeCroy power supply were designed in by someone who did not read the datasheet.
- Hope my photo demonstrated it makes sense to open the power supply after some years and clean it.

Have another one: The 2x 330uF 63V caps in the additional modular 48V power supply are by no means sufficient for the rated 7.8 A output current. They only survive because that power supply runs at a mere 2.5 A. In our SDA 6020 i will replace them by polymer caps that are rated for ripple currents of 4 A each.

Regards, Dieter

[Converter]

This is the right decision. I'm, in my TDK-Lambda Vega power supplies, replaced absolutely all of them with polymer capacitors, since the original capacitors in this place die very quickly.

[Mechatrommer]

can anyone help post their DDA3000 (WavePro) calibration data here? i have here one DDA3000 with damaged HDD, lots of bad sectors, i think C: will be totally unreadable. i did manage to recover D: but there is no calibration folder in it, other than one calibration.zip file in Logs folder but thats dated 2008, the DSO was last calibrated 2016 but those data cant be found, if its in C: then its very little hope to recover that. the DSO was sold as working but when i received it, it did boot slowly to xstream SW splash screen and then halted there, i shut the power off and thats the first and last time it entered the OS. anyone have clue how much cost to get new HDD/OS install for this scope from official LeCroy service? i think i want to make partial refund request, but it should be reasonable, so i ask advice from you people.

[Converter]

I think that it is not rational to pay LeCroy for a new HDD with an installed X-Stream, while it is freely available. But in case of damage to the disk, you will have to perform a calibration procedure at a service center that can do it. I can send you the files from the calibration folder of my oscilloscopes, but this is an individual setting and is unlikely to be useful for your oscilloscope unit.

[Mechatrommer]

I think that it is not rational to pay LeCroy for a new HDD with an installed X-Stream, while it is freely available. But in case of damage to the disk, you will have to perform a calibration procedure at a service center that can do it. I can send you the files from the calibration folder of my oscilloscopes, but this is an individual setting and is unlikely to be useful for your oscilloscope unit.

yes you are right i forgot that we can rebuild the system, but the calibration data. so it will come down to the cost of calibration process. please post your calibration data for DDA3000 anyway here or in my PM, i'm still interested, if you have say 3 unit 3000 with different calibration data, if its not too difficult, i will be interested in all of them too. since i'll consider this DDA3000 unit as broken or faulty or calibration data missing, i think i want to play with different calibration data to see how signal will show up. i have calibration data for 5005 and 6000, but i'm afraid if some data are specific to WaveMaster unit that is not available in WavePro. i'm in the process rebuilding this unit including i'll open up the psu to see whats inside, since this unit is much dusty than my previous WM units.

[pquadrat]

I think the cal data is stored in flash also. There is a thread in this forum. And I am sure when I tested Win7 with my Wavemaster, it retrieved the cal data from there. But the volume "D:" with name USERDATA has to be there before You start the software the first time.

[Mechatrommer]

i'm not sure the cal data stored in flash is originally from factory (2003) or the latest calibrated value (2016), if its from factory 2003, then i already recovered version 2008. and one thing is, the dda3000 was working without cal data in USERDATA, so i guess the SW loaded the cal data from C: (now corrupted) iirc there's discussion for flash recovered cal data, it will be stored in C, inside the installed xstream directory iirc.

[Converter]

Here are copies of the Calibration folder from multiple oscilloscopes. 1 oscilloscope WP7300A, the rest - WP7200/7200A. I have no calibrations for DDA3000, it arrived without a hard disk.

https://drive.google.com/file/d/1RG1uCdbd8-i1jWuGvACRCZU--692ck3l/view?usp=sharing

[Mechatrommer]

Here are copies of the Calibration folder from multiple oscilloscopes. 1 oscilloscope WP7300A, the rest - WP7200/7200A. I have no calibrations for DDA3000, it arrived without a hard disk.
https://drive.google.com/file/d/1RG1uCdbd8-i1jWuGvACRCZU--692ck3l/view?usp=sharing

super thank you Converter! i hope this will help me in analyzing the effect of the displayed signal with different calibration parameter. can you tell more about your DDA3000 experience? are you using xtream flash recovered calibration data? or using data from different scope? we can say that my DDA3000 is also without hard disk (unrecoverable original data) mostly in D are some waveforms screenshots that are very close to useless.

[Converter]

I didn't spend too much time on my DDA3000. Just deployed the disk image from WP7300, and it seems I have not yet seen the error messages.
For quite a long time ago, I noticed that the calibration data of the input amplifiers EEPROMs automatically overwrites the some files in the Calibration folder in the event that their ID does not match.

[ollopa]

The calibration folder has the hidden attribute set so be sure you set "view hidden folders" in Windows or use the /a switch with dir in DOS.  If you have a D: drive with the proper volume label then it will be used for cal file storage.

[Mechatrommer]

be sure you set "view hidden folders" in Windows

already done since i touched a keyboard (anyone who have not set this mode should not poke this thread with 10 feet barge pole) not just to see calibration data, but to see if any virus misbehavior that antivirus cant detect. the DDA3000 HDD was initially unreadable both C and D, i have to use WinXP install disk to enter DOS mode in the DSO and use CHKDSK /p to recover the volumes. only D become readable again, C is still unreadable (Windows will ask to format this volume which i wont do) i guess too much bad or corrupted sectors on the FAT itself. just if low level sector by sector disk clone still exist...
ps: aha i found it! https://hddguru.com/software/HDD-Raw-Copy-Tool/ i hope it will works, need a good sleep first...

[Mechatrommer]

if anyone can givew some idea here... AFE issue on DDA3000... https://www.eevblog.com/forum/repair/gaining-access-to-the-acquisition-board-on-a-lecroy-wavepro-7200/msg2376651/#msg2376651

[Converter]

I do not know what I could advise in this case. This is just like a ball array chip dump, if the sensitivity to the signal is preserved. Try to do the heating, or rebowling. In addition, I would try to send messages to all LeCroy sellers who potentially traded parts on eBay - maybe they have other parts left.

Can someone say, the functionality of this part, that in the photo of the Mechatrommer's input amplifier (highlighted in red)? What are their characteristics?
In the second photo, a similar item from my oscilloscope.

[DaJMasta]

Looks like a small piece of semirigid coax, from the RL silkscreen labeling, it could be an inline relay that wasn't needed for the model or in final production that was just bridged with the coax.

[Converter]

An interesting opinion. I was just worried that this was some kind of special piezo filter, or delay line.
I'm trying to figure out what the difference between the input amplifier for the Lecroy 7200 models is from the 7300. Does this differ only in the software setting of the digital filter (DSP), or are there also differences in hardware?

[DaJMasta]

An interesting opinion. I was just worried that this was some kind of special piezo filter, or delay line.
I'm trying to figure out what the difference between the input amplifier for the Lecroy 7200 models is from the 7300. Does this differ only in the software setting of the digital filter (DSP), or are there also differences in hardware?

Hardware.  You don't overmake the bandwidth by 50% for software limitations, especially at the time of release, in an entirely modular architecture.  I can't find my images of a 7300A frontend, but the heatsink for the main amp comes out through the frontend cage, and I am pretty sure the physical layout is different as well.


I would be very, very surprised if you could take a WP7200's frontend and somehow software unlock it to 3GHz of bandwidth.  You can definitely get amplification to 3GHz and beyond, you can see some pretty quick signals on, but I don't think you can get nominal amplification even when characterized and corrected for (and you'd need to calibrate it) to 3GHz.

[Converter]

Boards amplifier-attenuator are of the same exterior design for WP7300 and WP7200.
I do not even see the difference when adding the index "A":




But changing the software key "MBW" to "HBW" for WP7200 does not change it in WP7300, if you check the bandwidth.

This changes the bandwidth, but to a very small extent.
In this picture you can see the bandwidth change in WP7300A depending on the enabled options: "LBW"(blue), "MBW"(yellow) and "HBW"(green), denotes WP7100, WP7200 and WP7300 respectively:


Same for WP7200:


Here is the graphical overlay of the WP7300 and WP7200 frequency response while the HBW option is enabled in both:


Similarly, with the DSP turned off:


I took these screenshots 2 years ago.
As you can see, there is something else besides software bandwidth settings.
However, I have not yet verified the effect of the calibration data that is stored in the EEPROM chip in each of these amplifier boards (which are duplicated in the Calibration folder). I also see several trimmers on this board that can be explored.




Addition.

See what is contained in the EEPROM 24LC64 (Soic8) of the WP7300A entry boards (in one of the channels):