Author Topic: Tektronix THS7xxx Scope Hack/Teardown/Discussion - FW 1.16 found  (Read 144614 times)

0 Members and 2 Guests are viewing this topic.

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 5453
  • Country: de
zucca,
Should I send you my 720P, so you can read the FW out?
There are 3 kinds of people in this world, those who can count and those who can not.
 
The following users thanked this post: BravoV, Zucca

Offline ZuccaTopic starter

  • Supporter
  • ****
  • Posts: 4295
  • Country: it
  • EE meid in Itali
YES!!!
HV can read my mind.
I will send you my DMM7510 as a small thank you...

Let´s talk about this in May when I will be back home.
Can't know what you don't love. St. Augustine
Can't love what you don't know. Zucca
 

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 5453
  • Country: de
Great!
Let me know, when you are ready.

There are 3 kinds of people in this world, those who can count and those who can not.
 

Offline sky2city

  • Contributor
  • Posts: 34
  • Country: cn
I have one THS720A, with FW 1.11.  then I upgrade it to 720P, it work.
thanks you guys.

 

Offline grinch

  • Newbie
  • Posts: 4
hello again,

just changed old, crumbly molded foam insert in the case by a "new" one. Second hand packing foam was used. if someone needs sizes of original foam see images.

hope it helps.

« Last Edit: June 29, 2018, 05:07:44 pm by grinch »
 
The following users thanked this post: Zucca, walter.aranda

Offline ZuccaTopic starter

  • Supporter
  • ****
  • Posts: 4295
  • Country: it
  • EE meid in Itali
Nice!

Sorry guys, I got a girlfriend and all my planes are arse up now.
Can't know what you don't love. St. Augustine
Can't love what you don't know. Zucca
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 7547
  • Country: 00
  • +++ ATH1
YES!!!
HV can read my mind.
I will send you my DMM7510 as a small thank you...

Let´s talk about this in May when I will be back home.

Zucca, hows the progress on extracting the firmware ?

Offline ZuccaTopic starter

  • Supporter
  • ****
  • Posts: 4295
  • Country: it
  • EE meid in Itali
0%.

That girl and work are sucking all my time.
« Last Edit: June 30, 2018, 11:24:59 am by zucca »
Can't know what you don't love. St. Augustine
Can't love what you don't know. Zucca
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 7547
  • Country: 00
  • +++ ATH1
0%.

That girl and work is sucking all my time.

LOL .. ok , take your time, and TIA for doing this, also to HighVoltage, you guys rock.  :clap:  :-+

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 5453
  • Country: de
@zucca
I just came back from a long trip and can send it out any time now.
But you might need more time with the new GF.
Just let me know, when you are ready.
There are 3 kinds of people in this world, those who can count and those who can not.
 

Offline ZuccaTopic starter

  • Supporter
  • ****
  • Posts: 4295
  • Country: it
  • EE meid in Itali
HV

guess what, I am on business trip in Leipzig until end of August... so away from my Lab again for 2 months.  :'(
Can't know what you don't love. St. Augustine
Can't love what you don't know. Zucca
 

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 5453
  • Country: de
Ohh, you are busy guy these days.
No problem, just let me know, when time permits.
There are 3 kinds of people in this world, those who can count and those who can not.
 

Offline analogRF

  • Frequent Contributor
  • **
  • Posts: 955
  • Country: ca
Re: Tektronix THS7xxx Portable Scope Hack/Teardown/Discussion
« Reply #237 on: August 27, 2018, 03:22:11 pm »
When I received my (then) THS710A, both scope channels showed some DC offset that changed with the input sensitivity ranges selected. DC path compensation didn't correct this and I read somewhere in the forum that this is relateds to "worn out" analog optocouplers (CNR-201A) that are used to transfer the DC component of the input signal across the isolation barrier. Since these components are not particularly expensive. I decided to get some and replace them.

Guess what the result was: Same picture as before  :o

As a last resort I thought the complete factory calibration of the scope section might possibly help, and fortunately, it did the trick!  DC offset vanished completely. ;D

I just wanted to post the information in this thread since it seems to be the most comprehensive and as I understand, many of the THS700 models suffer more or less severely from the DC offset over the time. Calibration is very simple and all that's required for the job is a good (arbitrary) function generator that's capable of outputting a square wave with half-way decent rise time (<20ns) and a DC voltage (the accuracy of the latter should be checked with a reliable multimeter).  Fortunately, the scope section and the DMM section of the THS700 is calibrated separately and the DMM section calibration is very unlike to deteriorate over time.

All the information required for the job is available in the service manual, starting at page 51: http://de.tek.com/oscilloscope/ths720a-manual/ths710a-ths720a-ths730a-ths720p-service-manual

My recommendation in case of DC offset problems is to try this calibration before swapping out any components, especially since removal of the analog isolators is a real PITA....

Cheers,
Thomas


P.S. Some calibration steps really take some time and the instrument may appear like it's crashed -- just get yourself a cup of coffee (or whatever you feel like having at the time...) and wait. The screen contrast is set to 50% during calibration, on my THS700 this results in a rather dim screen (normally it's set at 65%), so don't worry if something like this happens.

Hi TurboTom
I understand this is a very old thread but I was hoping you or someone on this thread can help. I have a THS720A with the same DC offset issues. I have done the calibration as you suggested twice but the offset comes back after a short while. It is good right after calibration but in an hour or two it comes back again. Any suggestion? Should I go for replacing the optocouplers?

thanks
 

Offline ZuccaTopic starter

  • Supporter
  • ****
  • Posts: 4295
  • Country: it
  • EE meid in Itali
analogRF

I have also the DC offset problem. I still need to recalibrate the unit and see what happens.

Do you see it in both channels? If not swap the opto and see what happen.

Can't know what you don't love. St. Augustine
Can't love what you don't know. Zucca
 

Offline analogRF

  • Frequent Contributor
  • **
  • Posts: 955
  • Country: ca
no its just one channel. the other one is fine. I did the calibration twice and it was good for like 1-2 hours (even without using the scope for any measurements) and the offset came back.

Have you tried replacing the optocoupler? some people say it works and some have not been successful.
I am sure if it is replaced the unit definitely needs re-calibration and that should be the case with replacing any component in the signal path (otherwise it will probably show an offset and even a wrong gain) but re-calibration without replacing the opto has not worked for me so far.

Is there any other component that might be the suspect?
 

Offline ZuccaTopic starter

  • Supporter
  • ****
  • Posts: 4295
  • Country: it
  • EE meid in Itali
No, I still need to calibrate the unit and see what happen then.

Mark the suspected opto with a sharpie, desolder both of them.
If you are in the mood you can test them in a separate board outside the device, just to do some sanity check.

If nothing comes out and swap them out and do the calibration again. You will see if the problem stays with the bad opto or not.

Are the power supply rails good on the bad channel?

Even if I tell you my suspects list you NEED to make some measurement and investigate yourself where the problem is.
Can't know what you don't love. St. Augustine
Can't love what you don't know. Zucca
 

Offline TurboTom

  • Super Contributor
  • ***
  • Posts: 1388
  • Country: de
Re: Tektronix THS7xxx Portable Scope Hack/Teardown/Discussion
« Reply #241 on: August 29, 2018, 08:56:28 am »
Hi TurboTom
I understand this is a very old thread but I was hoping you or someone on this thread can help. I have a THS720A with the same DC offset issues. I have done the calibration as you suggested twice but the offset comes back after a short while. It is good right after calibration but in an hour or two it comes back again. Any suggestion? Should I go for replacing the optocouplers?

thanks

Just pulled my old THS710 "+" from the shelf where it collects dust (...) and powered it up. Surprise, surprise -- seems like there's no offset drift. Right after turning on, both traces were within +- 0.2 DIV center (as far as it's visible on that crappy screen) when zapping through the ranges. ANd they also stayed there after some prolonged time of operation. I've got to add, I power the instrument from a switcher wall-wart with a fairly stable output, so battery drain or so isn't taken into consideration here. And as reported, I replaced all the analog optocouplers in the isolated frontends. If that made a difference for the long-time stability of mine -- my guess is as good as yours.

Since the optocouplers aren't that expensive, it may be worth a try if you're good at soldewring in the confinded space next to the shielding boxes... Would I do it again if I had to?  -Probably not, too many scoped around (bad case of TEA... ) and the THS700 really looks old-fashioned meanwhile. The power analysis function available after the hack may be useful once in a while tough.

Cheers,
Thomas
 
The following users thanked this post: analogRF

Offline wingerr

  • Newbie
  • Posts: 5
  • Country: us
Are there any pictures of the opto isolator in its natural habitat?
I'm going to go fiddle with it as my channel 1 is reading several negative kV. Don't know if there's a layout diagram available.
My firmware is v1.06, must have gotten one of the earliest ones.
 

Offline TurboTom

  • Super Contributor
  • ***
  • Posts: 1388
  • Country: de
@wingerr: Zucca posted a nice high-res photo of his EPROM readout setup a few pages ago that you can magnify and cut to get the board almost full-screen. You see the analog couplers (the four wider ones -- HCPL-201A) directly next to the input amp shielding cans. I wasn't able to remove the cans for optocoupler replacement so soldering in the new ones is quite a bit of surgery... IIRC, I cut the old ones out, removed the terminals individually, cleaned the pads and then swapped in the new ones. You will need the gull-winged versions.

 
The following users thanked this post: analogRF

Offline analogRF

  • Frequent Contributor
  • **
  • Posts: 955
  • Country: ca
why those cans/shields could not be taken out? 
Are the components inside them soldered directly on the main pcb and the cans are just shields? or they are completely shielded hybrid modules with separate pcb inside and only some pins coming out?

 

Offline TurboTom

  • Super Contributor
  • ***
  • Posts: 1388
  • Country: de
There are just too many substantial ground tabs soldered in the PCB. The PCB is a multilayer with a lot of copper for grounding in this area and I just didn't want to risk damaging it. With some pre-heating it's surely possible to get the cans out. It's just a question which way is less tedious. For my own part I decided that I'll be able to replace the couplers without removing the cans which (for me) turned out to be the right decision.
 

Offline Jim Narem

  • Contributor
  • Posts: 16
  • Country: us
Picture of an attenuator with the shield removed is here:  http://w140.com/tekwiki/wiki/File:Ths720std-scopeanalog.JPG.  There are other internal pictures of my unit (which is a STD, not an A) on the primary tekwiki page at http://w140.com/tekwiki/wiki/THS720

I used metcal hot tweezers with wide tips to do mine, another method is to use chip-quik with a hot air gun.  Clipping the pins is probably the easiest solution if you only have a soldering iron.

Usually only the output optoisolators (U45, U46) need to be replaced.  The two others (U50, U63) are reversed, they supply the attenuator with an isolated offset signal generated by the  DAC multiplexer U5.  They will all eventually wear out due to the ageing  characteristics of the LEDs in the optocouplers.  Not sure how good the long term supply of these parts will be due to the HP/Avago/Broadcom saga.
 
The following users thanked this post: walter.aranda

Offline analogRF

  • Frequent Contributor
  • **
  • Posts: 955
  • Country: ca
Picture of an attenuator with the shield removed is here:  http://w140.com/tekwiki/wiki/File:Ths720std-scopeanalog.JPG.  There are other internal pictures of my unit (which is a STD, not an A) on the primary tekwiki page at http://w140.com/tekwiki/wiki/THS720

I used metcal hot tweezers with wide tips to do mine, another method is to use chip-quik with a hot air gun.  Clipping the pins is probably the easiest solution if you only have a soldering iron.

Usually only the output optoisolators (U45, U46) need to be replaced.  The two others (U50, U63) are reversed, they supply the attenuator with an isolated offset signal generated by the  DAC multiplexer U5.  They will all eventually wear out due to the ageing  characteristics of the LEDs in the optocouplers.  Not sure how good the long term supply of these parts will be due to the HP/Avago/Broadcom saga.

Thanks for the great pictures. Actually I had seen and downloaded them from tekwiki long ago but somehow I had forgotten all about them.
Anyhow, I have a question about the offset issue although currently I am experiencing another problem on top of the offset issue which I'll explain in another post
I always thought U50 and U63 must be the culprits for the offset problem. Specially if the offset value changes with range, I suppose it must be going through the attenuators and therefore it must be coming from U50/U63. It seems to me if the offset is added by U46/U45 then you have a fixed amount of offset "divisions" on screen in all ranges. Am I wrong? At least in my case the offset value changes with range.
Now if my understanding is correct, I wonder why shouldn't we be looking at U53/U74 opamp? If that opamp has developed offset over time that can equally mess up the offset voltage the same way that U50/U63 do.
 

Offline analogRF

  • Frequent Contributor
  • **
  • Posts: 955
  • Country: ca
I did more experiments with my THS scope and now I can see I have two distinct issues which I have not seen discussed before. Hope someone could give me a clue as to how to go about repairing it.

1-The offset issue on channel 1: I have done the scope calibration twice and numerous SPC and the offset always "comes back" (read below) although it is much smaller than it used to be. Also the offset value changes with range.
However, yesterday I noticed something that must have been there all along. When I plug in the power adapter, the offset gradually, over 2-3 minutes, disappears! and when I unplug it, it takes 2-3 minutes for the offset to slowly come back. The batteries are brand new and fully charged and their voltage is solid. How can that be?

2-"High pass" type of response on channel 1: the bigger issue that I have with the scope is the freq response of channel 1. I connected my function gen with square waveform and a piece of coax with no termination directly to the scope. On channel 2, everything is honky dory on all ranges, the top/bottom of the pulses are flat and the gain is accurate and when I switch to sinusoidal the gain is perfect on all ranges. However, on channel 1 with the same setup, "only" on the two ranges (100mV-500mv and 10V-50V) the top and bottom of square wave is exponential with overshoot. When the wave period is long enough the voltage settles to the correct value, so the DC gain is OK (I have checked with DC voltage and the gain is perfectly ok) but if I increase the frequency it does not have enough time to settle so the square wave appears to have a larger amplitude than it should. And if I inject sine wave the gain is clearly wrong (larger than it must be) on those two ranges.

I tried to measure the input impedance of the scope ports by connecting my LCR meter. On the good channel I get something around 28.05-28.3pF on all ranges (it changes about 0.25-0.3pF when jumping between the ranges) but on the bad channel I get 30.9pF on the two "bad" ranges and 28.08pF on the two good ranges. The scope spec is 25+/-2pF and I had to insert a BNC-SMA adapter in order to connect my LCR meter tweezers, so I know 1-2pF error must be ok but I cannot explain the inconsistency on those two bad ranges on channel 1

One thing to note is that in those two ranges the X10 attenuator kicks in while in other ranges either there is no attenuator or only the X100 attenuator is in the signal path.

Any suggestions?
 
The following users thanked this post: walter.aranda

Offline Jim Narem

  • Contributor
  • Posts: 16
  • Country: us
I don't fully understand the operation of the scope, and I did this back in 2014, but when I instrumented the board and ran the cal proceedure it seemed that the signal passed IN to the atten board (called POS1 and POS2 in the CLIP schematic, generated by the DAC and passed by U50 and U63) changed the feedback loop on the optocoupler  going OUT of the atten board.  You need to read the datasheet on the HCNR201-300E optocoupler to see how the second photodiode is used to linearize the system.  After I replaced the OUT optocoupler, the cal process passed.  I watched the cal process by using a logic analyzer on the DAC to trigger a scope looking at the POS signal and the feedback line on the OUT optocoupler.  With that setup, I could see the cal stepping the POS signal voltage and failing out due to lack of range.  It may be possible that replacing either of the optocouplers may fix the problem; again, eventually both will have to be replaced due to ageing of the LEDs.

You can run the board outside the case to probe around, see my picture here:  http://w140.com/tekwiki/wiki/File:Ths720std-runningboard.jpg.  A differential probe is needed to see the useful signals.  I think I used a peltola to SMA adaptor to feed signals in to the atten board.

My scope always takes some time to warm up, initially with some negative offset, then I do an SPC and it's fine.

Your second problem sounds like a failing relay in the attenuator.  The RF cover isn't too hard to remove with chip-quik and a solder sucker and the relays were available last I checked in 2014.  The relays and the AC coupling cap are really the only things that mere mortals can repair on that atten board, so it's worth a try.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf