Tek THS720A thinks there are 110V on the input but there is just a 50 Ohm R

[Zucca]

Just got a THS720A from ebay, sold as "used, need calibration". I got some money back since I told the buyer I am 99% sure the unit is broken.

Anyway this is the situation, two 50 Ohm terminations on the inputs:



Here the funny CH1 500V/Div DC:



110.6V mean...  do we really think a calibration will compensate that ??

Then I found this:

http://www.tek.com/support/faqs/my-scope-has-developed-dc-offset-how-can-i-remove-it

I tried that, it pass the SPC (like the self test on startup)



but the offset is still there.

here the 500V/Div GND:



According to my investigation this offset is there also in the other ranges





and here the Probe Comp signal (of course without the 50R termination):

.

The good news is that CH2 looks reasonable:



so I "just" have to compare the CH1 circuit with the CH2 and find the bad guy.

Do you think I will be able to fix it? Any suggestions?

[tautech]

As it seems like the offset is present with differing input attenuation settings.
I suspect that a portion of a resistor divider on the input has drifted or been damaged.

It should not be hard to check this with a good DMM as you have a good Ch 2 to compare to.

Check all attenuation settings for offset to be sure before you rip it apart.

[Zucca]

Thanks tautech, what you says sound reasonable. Yes, the offset is there in all the attenuations. So I tear it down the beast, see here:

https://www.eevblog.com/forum/testgear/tek-ths720a-portable-scope-teardowndiscussion/

No hope to find the schematics in the entire www.
EDIT 17 Feb 2016: Look for THS7XXAPDiag.pdf and you will find it

I found a service manual supplement which looks interesting:

Acquisition System

The acquisition system amplifies the input signals, samples them, converts them
to digital signals, and controls the acquisition process under direction of the
processor system. The acquisition system includes the trigger, acquisition timing,
and acquisition control circuitry.

Attenuators.  2-3 Circuitry in the attenuator selects the input coupling and
attenuation factor. The processor system controls the attenuators with a serial
interface as well as with DC control voltages.
The Main Board assembly contains two attenuator hybrids (AT1 and AT2). To
allow floating measurements, the attenuators are individually isolated from the
rest of the Main Board by an isolation interface. Power to the attenuators is
coupled through transformers. Opto-isolators couple control voltages and DC
control signals to each attenuator. Transformers couple the high-frequency signal
path from each attenuator to the main board. Opto-isolators couple the low-fre-
quency signal path from each attenuator to the main board.

Each attenuator hybrid contains two attenuators, an AC coupling capacitor, three
relays, three relay drivers, and a preamplifier. The AC/DC coupling relay couples
the output of the BNC to the other relays in the attenuator hybrid. For AC
signals, the AC/DC coupling relay inserts a coupling capacitor into the input
signal path. The second relay selects a X10 attenuation factor. The third relay
selects a X100 attenuation factor.

Sampler Driver.  4 The output of the attenuator drives the sampler driver (U44)
inputs. The sampler driver provides gain amplification, bandwidth limit filters,
and outputs for the sampler and trigger signal paths.

Daculator.  5 The daculator (U5) performs the function of sixteen independent
D/A converters. The daculator provides DC control voltages to the attenuator
hybrids and acquisition system. The CPU controls the daculator serially.

Sampler.  7 The sampler IC (U7) provides two complete acquisition channels. It
contains all of the functions associated with sampling, A/D conversion, trigger
placement, and time base control.
Using a 30.303 MHz clock, the Sampler samples the analog signals from the
Sampler driver at up to a 500 MS/s rate. It converts each sampled value into an
8-bit digital value. The A/D reference voltage is 2.5 V.
The sampler finds the approximate trigger position by counting clocks. At high
sample rates, it locates precise trigger position by acquiring the trigger signal and
interpolating it to find the zero crossing.

So my attach plan is to check the power supply voltages, found a good ground and start to check the Attenuators and the Sampler Driver if I can find them.

I really hope it is not simply a calibration thing.

[simingx]

I had a THS720 with this problem once, no amount of SPC would get rid of it. A calibration at Tek solved it but it came back eventually...

[qno]

Sorry for the late post.

I had the same problem. A calibration does the trick.
You can do a calibration by yourself.
You can find the calibration manual online.

This problem occurs when you replace the battery before switching off.

[macboy]

You can't complain about 3.2 V offset on a 500V/div range! Even if it uses a 10 bit ADC, that is one ADC count, remember that full scale is +/- 2000 volt! In other words, there is no offset, you are imagining a problem, at least for that one range. Does the SPC procedure reduce the offset in all other ranges to less than 1/10 of a division? (e.g. 5 V/div range, offset should be < +/-0.5 V). If so, then you are doing fine.

[Zucca]

This problem occurs when you replace the battery before switching off.

Are you sure? I looked at the input stage of the Scope, I was thinking that the optoisolator for the DC Signals (for the AC components you have the little signal trafo) was the bad guy here. It can well drift with the time.

My next step is also to do a calibration, and I was probably confident it would have solve the problem. Thanks for the confirmation.
My unit is still in a box... shame on my I have no f... time.

You can't complain about 3.2 V offset on a 500V/div range!

I have just 50 Ohm connected to the BNC,  and the scope is telling me 110V. It can´t be right.

[amyk]

110.6V
11.35V
1.115V
119.2mV

I suspect something closer to the ADC than the input stage... the offset seems to be constant at the ADC's input. If you can check the data coming out of the ADC, that might also determine if the problem is on the digital (stuck bit?) or analogue side.

[Zucca]

110.6V
11.35V
1.115V
119.2mV

I suspect something closer to the ADC than the input stage...

that's correct, but according to the schematics before the ADC and after the range selection circuit there is the DC opto-isolator. The input circuit I ment is not the hybrid PCB right after the BNC CH input.

Anyway congrats for fixing the problem!

[qno]

I did this a couple of years ago so i cannot remember the details.
I had a calibration manual from the net.
You have to open the THS and remove or place a jumper. (not shure)

Then you need a few DC voltages you can make with a dc power  supply (or 2)
I think it needed 60 volt or so so I put two in series and adjust them with your DVM to required voltage.

You also need a 1 kHz 300 mV square wave generator with a rise time < 3 ns or so (it is in the manual)
I made one from some fast gates and a resistor divider resulting in a 50 ohm load.
It did the trick. (I had another 1 G scope to verify this.)

I think the THS writes the offsets to an EEPROM or so during power down  and reads it @ start-up.
When removing the Battery while still on this data is damaged and some how the software offset calibration or gain restores to a factory default that is beyond the range of the offset compensation allowed by the SPC.
 Remember that the scope can be calibrated without a a screwdriver so some clever switching of gain and bandwidth compensation is done by the software.

After this mine worked ok. I compared it with a 1 GHz Wavesurfer. But that was a decade ago....

[mk_]

No hope to find the schematics in the entire www.

Search for THS7XXAPDiag.pdf and you will find it.

Anyway, PM sent as the file is to large (1.7MB) for attachment.

[Zucca]

Thanks mk_ I already got them, my previous post is now updated.

[walter.aranda]

I had the same problem, I solved it by calibrating it, maybe it's not the optocouplers or anything like that, maybe it's just calibrating the oscilloscope, note: during the calibration make sure that the channel input is X1