TDS 380 calibration error

[Dw11in]

Hello

I recently bought a used TDS 380 that is for the most part functional with the exception of some issues with the vertical scale and dc/ac coupling, I have attachments to show you what I mean. This is my first oscilliscope by the way.

Basically what is going on is the vertical scale seems to have group of bad spots affecting both channels, the the input signal seems to distort and drop out. When I roll the vertical scale past this bad section, either lower or higher, everything works fine again. In fact I don't know why I get a calibration error message because both channels seem to be measuring accurately. And to add, the dc coupling for channel 2 doesn't seem to work, I get a flat trace with either a ac or dc signal applied. If I apply enough ac signal into ch 2 on dc, I start to get a signal but a distorted one. Other than these issues I have, everything seems to work perfectly fine.

I am wondering if maybe I have a bad hybrid, maybe changing the capacitors will help, or perhaps I have bad chips on the main board which I am hoping not! What do you think is going on?

tds380_1  This is the scope itself
tds380_2  When I power on the scope all the lights come on as normal and I hear relays clicking
tds380_3  Every time this scope powers up I get this screen, calibration fail but at least power up passes
tds380_4  I take at the calibration screen to notice the signal path failed but all other tests passed
tds380_5  Taking a look at the error log I see calibration error relative amp gain, what does that mean? Also the memory read error is from the floppy drive
tds380_6  Both channels show normal traces with no sources connected
tds380_7  Both channels connected to the 5V probe pin, no problems here
tds380_8  Both channels seem to work fine on ac coupling
tds380_9  With both channels on dc coupling, ch 2 seems to have a dc coupling problem
tds380_10   Here I am feeding both channels a 1kHz square wave output from a sound card in the good portion of the vertical scale, everything is fine
tds380_11   When I adjust the vertical scale to the group of of bad settings, both channels seem to slightly drop out and show distortion
tds380_12 and tds380_13   With the vertical scale set in the bad region and both channels connected to 5V probe pin, we again see abnormal results (both channels ac coupling)

[lowimpedance]

Sounds like the accq. memory is suspect.
Not sure if this model used any SMD electrolytics which were a big problem around this time with TEK cro's (400 and 500 series etc).

See this thread .
https://www.eevblog.com/forum/repair/pls-help-me-repair-tektronix-tds-380/
If not allready !!.

[Dw11in]

Thanks for the reply lowimpedance

I did a diag check to compare my results as you can see here, everything seems to pass except calibration. SMD caps, I don't think I seen any of these in this scope

I however notice ch1 ch2 bnc inputs seem to be slightly offset from the center as if maybe the previous owner might have been probing and possibly tripped across the cables, bending them slightly. When I saw this I opened the scope up and expected around the hybrid where the bnc's go into but didn't notice any cracks, broken, or lifted pins, looked ok . Also the external trigger had electrical tape wrapped around the bnc connector housing for some reason so that looked suspicious as well but appears to work fine

[siggi]

I'd guess you have dicky relays on the input attenuator boards. I couldn't quickly find a schematic of the attenuators, but you should be able to debug this two ways.
1. By ear: if you hear a relay click as you switch vertical attenuation level where the signal goes wonky, that's a tell.
2. By DMM: in DC coupling, the resistance from center of BNC to ground should be 1MOhm. If it deviates at any vertical level, that's likely a bad relay.


Sent from my iPad using Tapatalk

[Dw11in]

siggi, I did the resistance test on both channels dc coupled from center of bnc to ground.

Ch1 is 2 mega ohms, off by 1 Mohm
Ch2 is completely open circuit, I think I know why I can't measure dc voltage now!

Looks like I am making some progress now

[siggi]

The attenuator relays are a common fault on the TDS5/6/7, but the attenuator hybrids look very different from yours. You should be able to buzz the relays out now that you have access to them.

Also beware that leakage currents in the inputs can skew your resistance measurement. Measure in both directions to be sure that's not a significant error term.

Good luck!


Sent from my iPad using Tapatalk

[Dw11in]

When you say buzz out the relays, do you mean energize them? And to energize them my understanding is to power on the scope and cycle the relays while doing resistance measurements on the relay pins, correct? And to note I have 9V digital voltmeter to be used for checking the resistance, when I measured the resistance of ch1 from the center to ground of the bnc I got a reading on the scope of about ~2-3 volts dc. Would this dc voltage hurt anything as I check the relays given how sensitive the inputs are down in the mili volts

In case I do have a bad ASW5K relay, I found a replacement AW5K relay which seems to be identical and I think it should work. I included attachments for this information

[siggi]

Buzz out, as in measure the resistance through the contacts that should be closed. The DC measurement is normal, and you're not going to hurt the scope inputs with any decent DMM in resistance mode, as it will be current limited. You might get low resistance readings if the overload protection kicks in.
To be honest, though, I don't know the first thing about this scope's front end, except that it's different from the ones I know of .

Looks like 3 relays? Does one do the AC/DC coupling? When do the others toggle?

If you can find a schematic of the attenuators, post a link and it'll all be easier from there.

Sent from my Nexus 5X using Tapatalk

[SaabFAN]

Could the Cal-Error be that the Memory-Battery died and the SRAM holding the Calibration-Data has lost all it's contents?

I also read that there's a switch on these models somewhere at the back that enters some kind of factory-calibration mode, which allows the user to put in a fresh set of calibration-data.

Have you checked those possibilities? Might avoid days of troubleshooting if the rest of the device is perfectly fine - I still would try to replace as many of the capacitors as possible though.

[Dw11in]

siggi, I have the information you asked for in the attachments. Still working on resistance checking ch2 relays, difficult, voltmeter probe on bnc center and other on relay pins is this the correct way to do it?

SaabFAN, there is a jumper on the main board to unlock  calibration mode but the manual states I need special equipment that I don't have on hand. It could very well be that the battery did die as the error log shows this was recent. Maybe this scope sat a very long time and the battery died. If that is the case I'm wondering where the battery would be on this thing to change it.

[siggi]

On the TDS5/6/7 input attenuator relay failure will often lead to the so-called signal path compensation to fail. This is a self calibration sequence that's operator initiated. The 380 seems to have something similar, though whether it's a power-on confidence check, or operator initiated I don't know. The relay configuration is very different from the 5/6/700 scopes, if the service manual is to be believed:

"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 generates a calibration or ground signal. The third relay selects the attenuation factor (X1, X10, or X100)."

It's possible that OP may need to deal with NVRAM failure and calibration down the line. It also seems certain that this is futile until and unless the inputs work.


Sent from my iPad using Tapatalk

[siggi]

siggi, I have the information you asked for in the attachments. Still working on resistance checking ch2 relays, difficult, voltmeter probe on bnc center and other on relay pins is this the correct way to do it?

I would measure pin to pin on the relays. On at least one attenuation setting your input square wave becomes a sharp spike. I've seen this happen with dicky contacts, as they make like a small value capacitor instead of a short.

On the bad attenuator relays in my TDS784D, I measured tens of ohms contact resistance instead of milliohms. This measured on the NC contact as well as on the NO contact. The latter I couldn't measure until I had the relay out of circuit.

Are the attenuators soldered in?

The calibration on this scope is not necessarily a big ordeal, though I'm wholly ignorant about the particulars. On my 2430 it was a matter of supplying stable and accurate DC voltages at the scope's instruction.
There may also be other steps, involving sine waves of a particular frequency.
I'd fix the known problem first before worrying about that.


Sent from my iPad using Tapatalk

[siggi]

From the data sheet, it seems each relay has two independent sets of coils and NC/NO contacts. I wonder if they're all wired up and independent. I only see three external relay driver transistors.

In your shoes I'd measure the resistance between common and NC contacts on all relays. Then test for connectivity across them, e.g. Pin 1-10, 2-9 and so on.
After this measure the voltage on pins 1&10 (should be at ground), and lastly on 5&6 when the relays are on, or supposed to be on. Should be -5V when on.

Compare channels, as one possibility is that you have dead relay drivers...


Sent from my iPad using Tapatalk

[siggi]

Forgot to mention, measure resistance from common to NO when on for each.

This should narrow down the culprit, or possibly exonerate the relays.


Sent from my iPad using Tapatalk

[siggi]

Lastly, you may want to consider moseying down to the TekScopes yahoo group. You may get moar and better advice there


Sent from my iPad using Tapatalk

[Dw11in]

Thanks for the replies, checking pinouts of the relays now. So far what I noticed is all 3 relays on channel 2 seem to be making good closure contacts at the pins though when open some relays show 2 mega ohms and that is also accompanied with a scratchy feedback on the screen scope. This is my initial testing so far, more results to come later

also..
Yes the attenuators are soldered in

[Dw11in]

This was tedious and time consuming but here are my results for testing the relays of ch1 and 2

Almost forgot, I included tds380_relay layout.jpg to give an idea of the order that I tested these relays, if you look at the picture, starting from the top we have the front of the scope and the BNC inputs. Each relay I tested for each channel goes from relay 1,2, and 3, where relay 1 is closest to the front of the scope and the 3rd is the farthest away towards the Maxtex chips

[siggi]

This was tedious and time consuming but here are my results for testing the relays of ch1 and 2

Well then - so much for my theory of dicky relays :/.
I can't see any problem with those measurements, and I'm guessing 0.2Ohm is your DMM's lead resistance.
Does the scope still behave the same way? Some relay contacts are allegedly self-cleaning, and may come good with usage, though I've never seen this happen.

I don't get it - can you buzz out the signal path from the BNC input through the relays? E.g. the BNC input should have connectivity with some pin or pins on the first relay. That relay ought to be bridging over the DC block capacitor when either closed or open. I'm guessing the DC block capacitor is to the right of the topmost relay on each channel.

From there you should be able to trace the path of connectivity through the other two relays.

Another thing you can do is to trace the signal, see where it starts going sideways. Best to do this with another scope, but you can also do it by feeding the calibrator into your bad channel, then tracing the signal with a probe on the good channel.

[Dw11in]

I ran the 5V probe signal into channel 2 on AC coupling and probed with channel 1 noticing the signal  passing through all 3 relays just fine. When I adjusted the vertical scale on ch2 where distortion occurs I could still see a clear square wave on channel 1. I didn't notice any amplitude changes on ch1 has I changed the scale on ch2, leading me to believe that maybe the relays are just routing the signal so I would need to find the final output where I see those changes take place.

With ch2 on dc coupling I could immediately see the distortion on ch1 when probing the relay pins, I guess I have a dc coupling problem on the ch2 attenuator

The scope still behaves as it did before and I noticed now that the relays don't actually have to click when distortion occurs for either channel and to add it happens for both channels at the same time the same way! Here is an illustration to show you what I mean

Channel 1 vertical scale
Max V/div                        (*)attn relay click                               Min V/div
    <-------------------------------*---------------------------------->
                                       | | | | | | | | | | | | |
                                Distortion appears in this area

Channel 2 vertical scale
Max V/div                        (*)attn relay click                               Min V/div
    <-------------------------------*---------------------------------->
                                       | | | | | | | | | | | | |
                                Distortion appears in this area

I hope this illustration gives you an idea what is going on here. I am starting to believe that along with the relay that attenuates the signal, it seems like some where there is solid state / digital or computerized attenuation happening here, reducing or increasing the signal, and I think where ever that is happening the problem might be occurring there. Perhaps a bad capacitor could fix this, my thoughts anyways

OPS! I played with the scope some more and I have revised illustrations, sorry!

Channel 1 vertical scale
Max V/div                        (*)attn relay click                               Min V/div
    <-------------------------------*---------------------------------->
                                    | | | | | | | |
                                Distortion appears in this area

Channel 2 vertical scale
Max V/div                        (*)attn relay click                               Min V/div
    <-------------------------------*---------------------------------->
                                    | | | | | | | |
                                Distortion appears in this area

[siggi]

Nice, two problems interacting.
This makes sense, as I looked at your initial pics again, and I can't fit them to relay trouble alone.
The attenuators have variable gain, pin 23 on each is probably used to control gain in between relay clicks as well as for the vertical vernier control.
You should see the voltage jumping up and down as you change the vertical control. You may also try to play with the vertical vernier, see if that helps.

Looks like you'll have to recalibrate, but for that to succeed, you have to fix the DC coupling issue. Incidentally, on other Tek attenuators I'm familiar with, the relay simply bridges over the coupling cap.
According to your measurements, the relay is good, and the coupling cap looks good as AC coupling works.
If all of this is true, then the relay has to be somehow shorting the signal to GND or some low impedance point in DC coupling.
Also your input impedance measures 2M, which is off for sure.
I can explain neither of those - is there any chance of measurement error?

Note that on my 2430 there's an option under the calibration menus to "cold start" or some such. This re-initialized the calibration constants to defaults. You might consider hunting for that option on your scope, as it should put it nearer to sane.
If this helps, then you know you'll need new NVRAM or battery, and you'll have a picture of what else (if anything) is sideways.


Sent from my iPad using Tapatalk

[Dw11in]

Have an update here and also bare to mind I am a bit of a newbie when it comes to this electrical stuff though I am starting to learn and pick up on things... Anyways I figured out where the outputs are from the attenuators and how that goes to the sampling driver. Connecting CH2 to the 5V test probe point I probed with channel 1 on pins 29,28 and it was there that I seen the square wave on channel 1, when I adjusted the vertical scale on channel 2 to make it mess up I saw the exact same results probing on channel 1. Reversing the operation and probing with ch2 on the output pins of ch1, I adjusted ch1 where it would distort the square wave and got the same results! So it seems now that I think I isolated the fault to both attenuators which is strange. I feel like I want to say now that some circuit connected to both attenuators  might be sending an abnormal signal at some point, causing the problem.

Also, I probed the variable gain adjustment of ch2 with ch1 and I do get a voltage change when adjusting the vertical scale. With the vertical scale turned to the max volts/div I read -.92 volts and when I click one step down I get -1.82 and from there it just stays at -1.82 all the way to the minimum vertical scale, and the relay clicking seems to have no effect. Tracing where this voltage is coming from, the daculator, I found that and found the pins to probe to get that voltage so now it seems I traced the voltage gain adjustment source. As for the voltage only dropping one time, I am not sure why it does this, maybe there are more pins to factor in for this to make sense.

You mentioned vertical vernier, what is that?

[siggi]

I feel like I want to say now that some circuit connected to both attenuators  might be sending an abnormal signal at some point, causing the problem.

I think there is unfortunately a possibility that your attenuators are toast. If they were presented with out-of-range signals at any point, that could explain why they're both showing garbage at the same attenuation range.
Without schematics for the attenuators themselves (or an attenuator to play with), I can't help you with the diagnosis.
Do you see any markings on the scope that'd indicate where it came from? A common student schenario is: input a terminal signal to a channel, see nothing sensible on the screen, try the other channel.
Now both channels are toast on the afflicted range :/.

Also, I probed the variable gain adjustment of ch2 with ch1 and I do get a voltage change when adjusting the vertical scale.

This signal should be DC - better to measure it with your DMM.

With the vertical scale turned to the max volts/div I read -.92 volts and when I click one step down I get -1.82 and from there it just stays at -1.82 all the way to the minimum vertical scale, and the relay clicking seems to have no effect. Tracing where this voltage is coming from, the daculator, I found that and found the pins to probe to get that voltage so now it seems I traced the voltage gain adjustment source. As for the voltage only dropping one time, I am not sure why it does this, maybe there are more pins to factor in for this to make sense.

There is a serial data input to the attenuator, likely this picks the coarse attenuation or gain, and the gain adjust then is used to tweak it to a calibrated value.

You mentioned vertical vernier, what is that?

This will be in the vertical menu, could be called something like "fine control" or some such. This allows you to continuously tweak the gain of the pre-amp. As you play with the vernier, you should see a change in the daculator output signal.
This is per-channel, so you should be able to play with the gain adjust signal to each channel individually.

On old style analog scopes there'd be a potentiometer that was latched when in calibrated deflection, and you'd turn it to tweak the gain to e.g. get a signal to line up with graticule markers for rise time measurement and such. See e.g. this video https://youtu.be/eRA1Mx506fw?t=310.

[siggi]

Actually, the fact that you don't see the "distortion" in two ranges gives some hope that this is settings-related rather than physical damage.
The way the attenuators and pre-amps work on other Tek scopes I'm familiar with, is that there are "hard" attenuation steps done by relay. Between the hard attenuation steps, there are "soft" steps that pick either gain or attenuation (I don't know which) in a 1-2-5 sequence. These gain/attenuation steps are re-used within each hard step, so you could get e.g. {1,2,5}/100, {1,2,5}/10 and {1,2,5}/1.
At the very highest sensitivities (say 1, 2, 5mV/div), there will be "soft" gain steps that are only used down at the bottom of the range.

So, all this to say that it seems your attenuators have only 2 hard attenuation steps. If the attenuators are bad, you'd therefore expect to see the "distortion" you see on the upper attenuation range to be reflected in the lower attenuation range. You would probably see this a few steps up from the highest sensitivity.

It might help if you report the number of steps above and below the attenuator relay click, as well as the level where the relay clicks...

Also, as this may get fairly involved, I suggest (again) you take this over to the TekScopes Yahoo forum https://groups.yahoo.com/neo/groups/TekScopes/info  where the Tek greybeards hang out.

[siggi]

I've just looked through your photos again, and I don't understand where the "bad range" is.
In pics _7 and _8 you show both channels 5V&2V/AC working fine.
In _9 you show CH2 5V/DC bad.
In _10 you show both 1V/AC working fine.
In _11 both channels at 500mV/AC bad.
But then _12 and _13 show 2V/AC is bad in a different mode.

You also don't show how you're connecting the signal in, except in the last two photos. Note that your scope has probe sensing, and it looks like you're using a 10X Tek probe in the last two photos. Did you hook up the signal the same way for all photos?

This is important, for our purposes, because when you connect a probe with the identification pin, the scope will adjust the on-screen display for the probe attenuation, so e.g. 200mV/div will become 2V/div.

[Dw11in]

Siggi, I included an attachment to address your concerns trying to interpret the pictures above, sorry about that. tds380_v scale demonstration.jpg demonstrates the problems I am having with the vertical scale of the scope. In this test I am running a 1kHz square wave out of an iPad into RCA-BNC adapters of both channels. Coupling is set to AC for both channels as well as 20 MHz bandwidth limiting. I could have used the probe test point but due to the fact that the signal would be to strong at the lower levels, instead I took this approach and when the signal appeared to be getting large on the scope's channels I would turn the iPad's volume down to compensate. This setup might not be the best but it got the job done.

Also I joined the tek yahoo group, just awaiting a response at this moment. Below you requested more info about the v scale increments and when the attenuator relay clicks... Please note I did not have the 10x Tek probe attached during this check

(- - -) Soft attenuation and vernier control (***) Hard attenuation relay click (\\\) Distortion
 10V     5V     2V     1V     .5V     .2V     .1V     .05V     .02V     .01V     .005V     .002V
|                                \\\\\\\\\\\\\\\\\\\\                                                                     |
| - - - - - - - - - - - - - - - - - - - - - - - -*** - - - - - - - - - - - - - - - - - - - - - - - - - -|
|   With vernier control distortion starts at 990mV and ends at 100mV  precisely        |
    Turing v scale knob, increments stop as seen above 10V 5V...

I will be testing the voltage to attenuator gain settings soon so those results to come for channels 1 and 2.

And one last thing, thank you very much for all your help! 

*Relay clicks at exactly .1V or 100mV